STML 37 Prev

Chapter 1
Arithmetic of the p-adic

Numbers
The aim of the first chapter of this book is to introduce its main
protagonist: the field of p-adic numbers Qp , defined for any prime p.
Just like the field of real numbers R, the field Qp can be con-
structed from the rational numbers Q as its completion with respect
to a certain norm. This norm depends on the prime number p and
differs drastically from the standard Euclidean norm used to define
R. Nevertheless, in each of the two cases, completion yields a normed
field (R and Qp ), and this general concept is studied in detail in §1.2.
But first (§1.1), we recall the completion procedure in the more fa-
miliar case of the reals (this takes us from Q to R), and only then do
we go on to its generalization to arbitrary normed fields (§1.3).
Putting these preliminaries aside, we come to the central section
of Chapter 1 (§1.4), where the construction of Qp is actually carried
out.
§§1.5–1.8 are devoted to the algebraic and structural properties
of the p-adic numbers. Here, as in subsequent parts, we will be con-
stantly comparing Qp and R, stressing both their similarities and their
differences. Finally, §§1.9 and 1.10 treat additional topics and are not
closely related to the rest of the book.
1
2 1. Arithmetic of the p-adic Numbers
1.1. From Q to R; the concept of completion

The real numbers, denoted by R, are obtained from the rationals by
a procedure called completion. This procedure can be applied to any
metric space, i.e., to a space M with a metric d on it. Recall that a
function
d:M ×M →R
defined on all ordered pairs (x, y) of elements of a nonempty set M
is said to be a metric if it possesses the following properties:
(1) d(x, y) ≥ 0; d(x, y) = 0 if and only if x = y;
(2) d(x, y) = d(y, x) ∀ x, y ∈ M ;
(3) d(x, y) ≤ d(x, z) + d(z, y) ∀ x, y, z ∈ M .
The function d is also called the distance function.
We say that a sequence {rn } in a metric space M is a Cauchy
sequence if for any ε > 0 there exists a positive integer N such that
n, m > N implies d(rn , rm ) < ε. If any Cauchy sequence in M has a
limit in M , then M is called a complete metric space.
Theorem 1.1 (Completion Theorem). Every metric space M can be

, D) such that
completed, i.e., there exists a metric space (M
is complete with respect to the metric D;
(1) M
contains a subset M
(2) M 0 isometric to M ;
(3) M (i.e., each point in M
0 is dense in M is a limit point for

M0 ).
The proof that can be found e.g. in [13, Theorem 76] consists
in an explicit construction of the completion M and the metric D
on it. We start with the collection {M } of all Cauchy sequences in
M , convergent or not, and turn it into a metric space. But first we
introduce an equivalence relation on {M }: two Cauchy sequences an
and bn are called equivalent, we write {an } ∼ {bn }, if d(an , bn ) → 0.
(It is easy to check that this is an equivalence relation on {M }.)
We define M to be the set of equivalence classes, M = {M }/ ∼.
The metric D between two equivalence classes of Cauchy sequences
1.1. From Q to R; the concept of completion 3
A = ({an }) and B = ({bn }) is defined by the formula
(1.1) D(A, B) = lim d(an , bn ).

n→∞
We leave it to the reader to check that the limit above always exists
and does not depend on the choice of representatives in the equiva-
lence classes (Exercise 7) and that D indeed is a metric on M (Exer-
cise 8). In §1.3 we will give the complete proof of this theorem in the
particular case of metric spaces called normed fields, which includes
Q.
The completion procedure applied to M = Q with the usual
Euclidean distance between rational numbers,
(1.2) d(r1 , r2 ) = |r1 − r2 |,
yields the real numbers R. Notice that this distance “came from” the
Euclidean norm on Q, which is the ordinary absolute value.
Another description of the completion of Q yielding R, more fa-
miliar and less sophisticated than the one above, is based on infinite
decimal fractions. Every positive real number a can be written as an
infinite decimal fraction
∞

(1.3) a= ak 10−k ,
k=m
where m is a certain integer and the coefficients or digits ak take the

values
0, 1, 2, 3, 4, 5, 6, 7, 8, 9.
This representation is unique unless ak = 0 for all k > n and an = 0,
in which case a has a second representation with ak = ak for k < n,
an = an − 1, and ak = 9 for all k > n. Conversely, any infinite
decimal fraction represents a point on the “number axis”; thus it is
convenient to identify real numbers with infinite decimal fractions.
If represented as infinite decimal fractions, rational numbers are
characterized by the property that they are eventually periodic (Exer-
cise 2). It is easy to construct a Cauchy sequence of rational numbers
which has no limit in Q:
.1, .1011, .10110111, .1011011101111, . . . ;

hence Q is not complete with respect to the Euclidean distance. On

the other hand, any equivalence class of Cauchy sequences of rational
numbers has a representative which is a sequence of partial sums of
a series of the form (1.3) whose limit is an infinite decimal fraction
(Exercise 1), and the same is true for any Cauchy sequence of infinite
decimal fractions. In other words, the set of real numbers is com-
plete with respect to the Euclidean distance (Exercise 3), and the
construction of real numbers by means of infinite decimal fractions is
equivalent to the completion procedure with respect to the Euclidean
distance.
This representation can be generalized to the representation to
the base g, where g is an integer greater than or equal to 2 and thus
∞

a= ak g −k ,
k=m
where the coefficients ak take values in the set {0, 1, . . . , g − 1}. Note
that the exponents −k of g are descending and tend to −∞.
The following notions can be defined in every metric space.
Definition 1.2. Let (M, d) be a metric space, and R+ denotes the

set of positive real numbers. The open ball of radius r ∈ R+ centered
at a ∈ M is the set
B(a, r) = {x ∈ M | d(a, x) < r}.
The closed ball of radius r ∈ R+ centered at a ∈ M is the set
B(a, r) = {x ∈ M | d(a, x) ≤ r}.
A set A ⊂ M is called open if for any x ∈ M there exists an open

ball B(a, r) ⊂ M containing x. A set A ⊂ M is called closed if its
complement M \ A is open.
Practically, completion of a metric space is often obtained by a

different construction described below:
Proposition 1.3. Let M be a complete metric space and let X be a

subset of M . Then X is complete if and only if it is closed in M . In
particular, the closure of X in M can be taken as its completion.
Exercises 1–8 5
Example 1.4. The completion of an open interval (a, b) with respect

to the usual Euclidean distance is the segment [a, b], the closure of
(a, b) in R.
For other examples, see Exercise 5.
Exercises 1–8
Exercise 1. Prove that any Cauchy sequence of rational numbers

with respect to the Euclidean distance has a representative which is
a sequence of partial sums of a series of the form (1.3).
Exercise 2. Prove that a number is rational if and only if its repre-

sentation by an infinite decimal fraction is eventually periodic.
Exercise 3. Use the representation of real numbers as infinite deci-

mal fractions to prove that the set of real numbers is complete with
respect to the Euclidean distance, i.e., that any Cauchy sequence of
real numbers has a limit.
Exercise 4. Prove Proposition 1.3.
Exercise 5. Prove that the following metric spaces are not complete,
and construct their completions:
(1) R with the distance d(x, y) = | arctan x − arctan y|;
(2) R with the distance d(x, y) = |ex − ey |.
Exercise 6. Prove that a metric space is complete if and only if the

intersection of every nested sequence of closed balls {Bn }, B1 ⊃ B2 ⊃
B3 ⊃ . . . whose radii approach zero consists of a single point.
Exercise 7. Let {an } and {bn } be two Cauchy sequences in a metric

space (M, d). Prove that the limit limn→∞ d(an , bn ) exists and does
not depend on the choice of representatives in the equivalence classes,
i.e., if {an } ∼ {an } and {bn } ∼ {bn }, then limn→∞ d(an , bn ) =
limn→∞ d(an , bn ).
.
Exercise 8. Prove that D defined in (1.1) is a metric on M
1.2. Normed fields

Both rational numbers and real numbers are prime examples of an
algebraic structure called a field. A field F is a set with two binary
operations usually called addition and multiplication which satisfy the
most basic properties of these two operations for numbers. Namely,
(1) ∀ a, b ∈ F , a + b = b + a (commutativity of addition),
(2) ∀ a, b, c ∈ F , a + (b + c) = (a + b) + c (associativity of
addition),
(3) ∃ 0 ∈ F such that ∀ a ∈ F , 0 + a = a (existence of zero),
(4) ∀ a ∈ F ∃ − a ∈ F such that a + (−a) = 0 (existence of the
additive inverse),
(5) ∀ a, b ∈ F , a · b = b · a (commutativity of multiplication),
(6) a · (b · c) = (a · b) · c (associativity of multiplication),
(7) ∃ 1 ∈ F such that ∀ a ∈ F × = F \ {0}, 1 · a = a (existence
of identity),
(8) ∀ a ∈ F × ∃ a−1 ∈ F × such that a · a−1 = 1 (existence of
the multiplicative inverse),
(9) ∀ a, b, c ∈ F , a · (b + c) = a · b + a · c (distributivity),
(10) 0 = 1.
An algebraic structure with only one binary operation satisfying the
properties (1) − (4) is called an abelian (or commutative) group. Cor-
respondingly, F with addition is called the additive group of the field
F , and F × with multiplication is called the multiplicative group of the
field F . An algebraic structure with two binary operations satisfying
the properties (1) − (6) and (9) is called a commutative ring.
An important property of a field is that it does not contain zero
divisors, i.e., a, b ∈ F × such that a · b = 0 (see Exercise 9).
Definition 1.5. Let F be a field. A norm on F is a map denoted
· from F to the nonnegative real numbers such that
(1) x = 0 if and only if x = 0;
(2) xy = x y ∀ x, y ∈ F ;
(3) x + y ≤ x + y ∀ x, y ∈ F (the triangle inequality).
1.2. Normed fields 7
The norm is called trivial if 0 = 0 and x = 1 for all x = 0.

Notice that, for any natural number n ∈ N, we have
. . + 1 ∈ F.
n · 1 := 1 + .
n times
We shall denote this element by the same symbol n as the correspond-

ing natural number.
Proposition 1.6. For any x, y ∈ F we have

(a) 1 = − 1 = 1;
(b) x = − x;
(c) x ± y ≥ | x − y |;
(d) x − y ≤ x + y;
(e) x/y = x/y;
(f) n ≤ n ∀ n ∈ N.
Proof.
(a) 1 = ± 1 · ±1 = ± 12 =⇒ ± 1 = 1.
(b) − x = (−1) · x = 1 · x.
(c) Follows from (b) and the triangle inequality for the norm
(see Exercise 10).
(d) Follows from (b) and the triangle inequality.
(e) Follows from (a) and the property (2) of the norm.
(f) Follows by induction from (a) and the triangle inequality.

Let d(x, y) = x − y. It follows immediately from Definition 1.5

and Proposition 1.6 that d is a distance function; indeed, Definition
1.5(1) implies that d(x, y) = 0 if and only if x = y, while Proposition
1.6(b) implies symmetry, and Proposition 1.6(d) yields the triangle
inequality. We say that this distance is induced by the norm · and
we will regard (F, · ) as a metric space.
Definition 1.7. A sequence {an } in F is said to be

• bounded if there is a constant C > 0 such that

an ≤ C ∀n;
• a null sequence if
lim an = 0,
n→∞
i.e., for any ε > 0 there is an N such that for all n > N
an < ε;
• a Cauchy sequence if
lim an − am = 0,
n,m→∞
i.e., for any ε > 0 there is an N such that for all n, m > N
we have an − am < ε;
• convergent to a ∈ F (we write a = limn→∞ an ) if
lim an − a = 0,
n→∞
i.e., for any ε > 0 there is an N such that for all n > N
an − a < ε.
It follows from the definition that any null sequence converges

to 0, and it follows from the triangle inequality that any converging
sequence is a Cauchy sequence: suppose limn→∞ an = a; then
ε ε
an − am = an − a + a − am ≤ an − a + a − am < + = ε
2 2
for n, m > N chosen for ε/2 in the definition of limit. In partic-
ular, every null sequence is a Cauchy sequence. Further properties
are listed below and are obtained by the same standard technique
(Exercise 11).
(1) Every Cauchy sequence is bounded.
(2) Let {an } be a Cauchy sequence and let {n1 , n2 , . . . } be an
increasing sequence of positive integers. If the subsequence
an1 , an2 , . . . is a null sequence, then {an } itself is a null
sequence.
(3) If {an } and {bn } are null sequences, so is {an ± bn }, and
if {an } is a null sequence and {bn } is a bounded sequence,
then {an bn } is a null sequence.
The following is a simple but very useful result.
Proposition 1.8. x < 1 if and only if limn→∞ xn = 0.
Proof. Let x < 1. Since xn = xn , we obtain
lim xn = 0,
n→∞
i.e., limn→∞ xn = 0. Conversely, if x ≥ 1, then for all positive n

we have xn ≥ 1, and therefore 0 = limn→∞ xn .
Definition 1.9. We say that two metrics d1 and d2 on F are equiv-

alent if a sequence is Cauchy with respect to d1 if and only if it is
Cauchy with respect to d2 . We say that two norms · 1 and · 2
are equivalent ( · 1 ∼ · 2 ) if they induce equivalent metrics.
Proposition 1.10. Let · 1 and · 2 be two norms on a field F .

Then · 1 ∼ · 2 if and only if there exists a positive real number α
such that
(1.4) x2 = xα
1, ∀ x ∈ F.
Proof. Suppose · 1 ∼ · 2 . If · 1 is trivial, then by Exercise 12

· 2 is also trivial, and hence (1.4) is satisfied for any α.
If · 1 is nontrivial, then we can choose an element a ∈ F such
that a1 = 1. Replacing a by a−1 if necessary, we can assume that
a1 < 1. Define
log a2
α= .
log a1
Notice that since the norms are equivalent, by Exercise 13 we have
a2 < 1 as well; hence both logarithms are negative and α > 0.
We will show that this α satisfies (1.4). First take x ∈ F with
x1 < 1; the cases x1 > 1 and x1 = 1 will follow from Exercise
13. Consider the set

(1.5) S = r = m/n | m, n ∈ N, xr1 < a1 .
For any r ∈ S we have
xm

xm
1 < a1 , so
n
n < 1.
a 1
Then by Exercise 13, xm

n < 1,
a 2
and so xm 2 < a2 , and x2 < a2 . The same argument holds
n r
with · 2 and · 1 interchanged, so we also find that

(1.6) S = {r = m/n | m, n ∈ N, xr2 < a2 }.
By taking logarithms, we can rewrite conditions (1.5) and (1.6) as
log a1 log a2
(1.7) r> , r>
log x1 log x2
since all logarithms involved are negative. But then we must have
log a1 log a2
= ,
log x1 log x2
because otherwise there would be a rational r between these two
numbers and only one of the conditions in (1.7) would be satisfied.
Therefore,
log x2 log a2
= = α,
log x1 log a1
and (1.4) follows.
Conversely, suppose x2 = xα 1 , and suppose {an } is a Cauchy
sequence with respect to the distance induced by ·1 . Given > 0, let
N be chosen for 1/α . Then for n, m > N we have xn − xm 1 < 1/α
and therefore xn − xm 2 < . The same argument holds with · 2
and · 1 interchanged, which concludes the proof.
Now we will describe all norms on Q equivalent to the absolute

value | · |.
Proposition 1.11. x = |x|α , α > 0, is a norm on Q if and only if
α ≤ 1. In that case it is equivalent to the norm | · |.
Proof. Suppose α ≤ 1. The first two properties of the norm are

obvious, so we only need to check the triangle inequality. Assume
that |y| ≤ |x|. Then

α
|y|
|x + y| ≤ (|x| + |y|) = |x| 1 +
α α α
|x|

|y| |y|α
≤ |x| 1 +
α
≤ |x| 1 + α = |x|α + |y|α .
α
|x| |x|
The first inequality follows from to the fact that tα ≤ t for t ≥ 1, and
the second because tα ≥ t for 0 ≤ t ≤ 1.
On the other hand, if α > 1, the triangle inequality is not satis-
fied: for example, |1 + 1|α = 2α > |1|α + |1|α = 2.
It will follow from Ostrowski’s Theorem (Theorem 1.50) that

Proposition 1.11 describes all norms on Q equivalent to the absolute
value | · |.
Definition 1.12. A norm is called non-Archimedean if it satisfies
the additional condition
(4) x + y ≤ max(x, y);
otherwise, we say that the norm is Archimedean.
Remark 1.13. The condition (4) of the norm implies the condition
(3), the triangle inequality, since max(x, y) does not exceed the
sum x + y. We will call this property the strong triangle inequal-
ity.
The metric induced by a non-Archimedean norm is said to be an

ultra-metric. Instead of the triangle inequality for the usual metric
d(x, z) ≤ d(x, y) + d(y, z),
it satisfies the strong triangle inequality
d(x, z) ≤ max(d(x, y), d(y, z)).
The corresponding metric spaces are called ultra-metric spaces.
The following theorem is a necessary and sufficient condition for
a norm to be non-Archimedean.
Proposition 1.14. The following statements are equivalent:
(a) · is non-Archimedean;
(b) n ≤ 1 for every integer n.
Proof. (a) ⇒ (b). We will prove this implication by induction.

Base of Induction. 1 = 1 ≤ 1.
Induction Step. Suppose that k ≤ 1 for all k ∈ {1, · · · , n − 1};
let us prove that n ≤ 1.
Observe that n = (n − 1) + 1 ≤ max{n − 1, 1} = 1.

From the inequality 1 = 1 ≤ 1 and the induction assumption,
we have n ≤ 1 for all n ∈ N. Since − n = n, we conclude that
n ≤ 1 for all integers n ∈ Z.
(b) ⇒ (a). We have

n

n k n−k
x + y = (x + y) =
n n
x y
k
k=0
n
n
n
≤ xk yn−k ≤ xk yn−k
k
k=0 k=0
≤ (n + 1)[max(x, y)]n .
So, for every integer n we have
√
x + y ≤ n n + 1 max(x, y).
Letting n tend to ∞, we obtain
x + y ≤ max(x, y).

Here we used both the fact that nk is an integer and the well-known
limit
√
lim n n + 1 = 1.
n→∞

This proposition helps explain the difference between Archimedean

and non-Archimedean norms. It can be restated as follows: a norm
is Archimedean if and only if it has the Archimedean property: given
x, y ∈ F , x = 0, there exists a positive integer n such that nx > y.
To see that, take x, y ∈ F with y > x. Then the Archimedean
property implies the existence of a positive integer n such that n >
y/x > 1, i.e., the norm is Archimedean. Conversely, if the norm
is Archimedean, there exists a positive integer n with n > 1. Then
nk → ∞ as k → ∞, and for some k, nk > y/x, which implies
the Archimedean property nk x > y.
It is easy to see that the Archimedean property is equivalent to
the assertion that there are integers with arbitrarily large norms:
(1.8) sup{n : n ∈ Z} = +∞.
We leave it to the reader to check that a norm is Archimedean if and

only if (1.8) is satisfied (Exercise 14).
The non-Archimedean property has other surprising implications.
Proposition 1.15. If the elements a, x of a non-Archimedean field
F satisfy the inequality x − a < a, then x = a.
Proof. By the strong triangle inequality,

x = x − a + a ≤ max(x − a, a) = a.
On the other hand,
a = a − x + x ≤ max(x − a, x).
Now x − a > x would imply a ≤ x − a, a contradiction.
Therefore x − a ≤ x, and a ≤ x. So, x = a. This
completes the proof.
Remark 1.16. This property can be restated in the following way:
for a, b in a non-Archimedean field F
a > b =⇒ a + b = a : the strongest wins.
Using the geometrical language, we can say: Any triangle in an ultra-
metric space is isosceles and the length of its base does not exceed the
lengths of the sides.
We leave the proof of the next rather surprising proposition to

the reader (Exercise 15).
Proposition 1.17. If · is non-Archimedean, then any point of an
open ball B(a, r) = {x : x − a < r} in F is its center, i.e., if b is in
B(a, r), then B(b, r) = B(a, r). The same is true for closed balls.
We shall conclude this section by showing that an Archimedean

norm and a non-Archimedean norm cannot be equivalent.
Proposition 1.18. Two equivalent norms ( · 1 ∼ · 2 ) on a field
F are either both non-Archimedean or both Archimedean.
Proof. It follows from Exercise 13 that if · 1 ∼ · 2 , then for

any integer n we have n1 > 1 if and only if n2 > 1. Hence by
Proposition 1.14 either both norms are non-Archimedean or both are
Archimedean.
Exercises 9–16
Exercise 9. Prove that a field does not contain zero divisors.
Exercise 10. From the triangle inequality for the norm on a field F
(Definition 1.5(3)) deduce that
|x − y| ≤ x ± y ∀x, y ∈ F.
Exercise 11. Prove that in a normed field the following assertions

hold:
(1) Every Cauchy sequence is bounded.
(2) Let {an } be a Cauchy sequence and let {n1 , n2 , . . . } be an in-
creasing sequence of positive integers. If the subsequence
an1 , an2 , . . .
is a null sequence, then {an } itself is a null sequence.
(3) If {an } and {bn } are null sequences, so is {an ± bn }, and if {an }
is a null sequence and {bn } is a bounded sequence, then {an bn }
is a null sequence.
(4) Let {an } be a Cauchy sequence, but not a null sequence. Prove
that there exist a number c > 0 and a positive integer N such
that for all n > N either an > c or an < −c.
Exercise 12. Prove that if · 1 ∼ · 2 and if · 1 is trivial, so is

· 2 .
Exercise 13. Prove that if · 1 ∼ · 2 , then x1 < 1 if and only

if x2 < 1, x1 > 1 if and only if x2 > 1, and x1 = 1 if and
only if x2 = 1.
Exercise 14. Prove that the norm · is Archimedean if and only

if
sup{n : n ∈ Z} = +∞.

1.3. Construction of the completion of a normed field 15
Exercise 16. Prove that if · is a non-Archimedean norm, then

· α is also a non-Archimedean norm for any α > 0. (Compare with
Proposition 1.11 for the Euclidean absolute value on Q.)
1.3. Construction of the completion of a normed

field
In this section, starting from an arbitrary normed field F (not nec-
essarily complete with respect to its norm · ), we will construct
another field, F, containing F , and supply it with a norm (induced
from the norm · of F ) in such a way that F will be a complete
normed field.
We have already seen (§1.1) that in the case of the rational num-
bers supplied with the ordinary (Euclidean) norm, the completion
procedure yields the reals R. The same procedure will be applied
later (see §1.4) to Q endowed with a completely different norm and
will yield the p-adic numbers. In the completion procedure, the main
role will be played by Cauchy sequences: it is equivalence classes of
Cauchy sequences from F that will be declared elements of the field F .
So we begin by discussing Cauchy sequences in an arbitrary normed
field.
Cauchy sequences can be added, subtracted and multiplied (Ex-
ercise 17), so the set of all Cauchy sequences in (F, · ), denoted
by {F }, becomes a commutative ring. Its identity element under
addition is the sequence

0 = {0, 0, 0, . . . },
and its identity element under multiplication is the sequence

1 = {1, 1, 1, . . . }.
It is clear that {F } is not a field since it contains zero divisors:
{1, 0, 0, . . . }{0, 1, 0, 0, . . . } =
0.
For every a ∈ F the constant sequence

a = {a, a, a . . . }
is Cauchy and therefore lies in {F }. Hence {F } contains a subring
isomorphic to F . Of particular importance is the set P of all null
sequences. We saw that P is a subset of {F }. In fact, P is an ideal

in {F } (i.e., a subring such that for all p ∈ P and all a ∈ F we have
ap ∈ P ). Indeed, if {an } and {bn } are in P , so is {an ± bn }, and if
{an } is in P and {bn } is a bounded sequence (in particular if it is
Cauchy), then {an bn } is in P (Exercise 11(3)).
Let F = {F }/P . Its elements are equivalence classes of Cauchy
sequences in (F, · ), two Cauchy sequences being equivalent if their
difference is a null sequence. Notice that constant sequences

a = {a, a, a, . . . },
where the a ∈ F belong to different equivalent classes in F for different

a. We shall denote the equivalence class of a Cauchy sequence {an }
by ({an }), so ({an }) is an element of F. We will think of F as a
subset of F, identifying a ∈ F with (
a) ∈ F.
Theorem 1.19. F is a field.
Proof. It is easy to check that F, with operations defined as follows:

if {an } ∈ A and {bn } ∈ B, then A + B = ({an + bn }) and A · B =
({an · bn }), is a commutative ring with the additive identity (
0) and
the multiplicative identity (
1). By Exercise 18, these operations do
not depend on the choice of representatives. Let us prove that F is a
field. Let A be an equivalence class in F different from the zero class
(
0) = P , and let {an } be any Cauchy sequence in A. By Exercise
11(4) there exist a positive number c and a positive integer N such
that
an > c ∀ n ≥ N.
Define a new sequence {a∗n } by

∗ 0 if 1 ≤ n ≤ N − 1,
an =
1/an if n ≥ N.
We claim that this is a Cauchy sequence. Indeed, if n, m ≥ N , then

1 1 am − an
∗ ∗
0 ≤ am − an = − = ≤ c−2 am − an ,
am an am · an
1.3. Construction of the completion of a normed field 17
and the claim follows since {an } is a Cauchy sequence. Let us denote
the equivalence class of the sequence {a∗n } by A−1 . Then
{an }{a∗n } = { 0, . . . , 0 , 1, 1, 1 . . . },

N −1 zeros
where the Cauchy sequence on the right differs from 1 by the null
sequence
{−1, . . . , −1, 0, 0, 0 . . . }.

N −1 (−1)’s
Thus AA −1
1), which proves that F is a field.
= (
Now we extend the norm · from F to F.
Definition 1.20. For any A ∈ F put

A = lim an ,
n→∞
where {an } is any Cauchy sequence in A.
In order to see that this norm is well defined, we must show that
the limit exists and does not depend on the choice of the Cauchy
sequence {an } in A. We have
|an − am | ≤ an − am
by Exercise 10, which implies that the sequence of real numbers
{an } is Cauchy with respect to the absolute value. Since the set of
real numbers R is complete, the limit defining · exists. Now take
a second sequence {an } ∈ A. By the same inequality we have
0 ≤ lim |an − an | ≤ lim an − an = 0;
n→∞ n→∞
hence limn→∞ an = limn→∞ an .
Proposition 1.21. · is a norm on F.
Proof. We must verify the three properties listed in Definition 1.5.

(1) If A = (
0), then {an } is a null sequence, and therefore A =
0. If A = ( 0) and A = ({an }), then there exist positive
numbers c and N such that for all n ≥ N we have an ≥
c > 0. Hence A > 0.
(2) Now let A = ({an }) and B = ({bn }). By the properties of

real limits,
AB = lim an bn = lim an bn
n→∞ n→∞
= lim an lim bn = AB.
n→∞ n→∞
(3) Similarly, we obtain A + B ≤ A + B.

Now we can define bounded, Cauchy, and null sequences in F

with respect to the norm · .
Theorem 1.22. F is complete with respect to the norm · , and F

is a dense subset of F.
Proof. We first prove the second part. Let A ∈ F, and let {am }
be a Cauchy sequence in F representing A. For each fixed positive
integer n, we consider the constant sequence an . Then the sequence
{am − an }∞m=1 represents A − (
an ), and since {am } is Cauchy, we can
write
(1.9) lim A − (
an ) = lim am − an = 0.
n→∞ n,m→∞
This proves that F is dense in F. Now suppose {An } = {A1 , A2 , . . . }

is a Cauchy sequence in F. By the density of F in F, for any An
there exists an element an ∈ F such that
1
(1.10) An − (
an ) < .
n
Therefore {An − (
an )} is null sequence, hence a Cauchy sequence in
F. We have
{(
an )} = {An } − {An − (
an )};
hence {(an )} is a Cauchy sequence in F , but since all its elements
belong to F , {an } itself is a Cauchy sequence in F . Let us denote the
equivalence class of {an } by A (in our notation, ({an }) = A). From
(1.9) and (1.10) it follows that {A − ( an )} and {An − (an )} are null
sequences in F , and hence their difference
{A − An } = {A − (
an )} − {An − (
an )}
1.4. The field of p-adic numbers Qp 19
is a null sequence in F. This implies that

lim A − An = 0,
n→∞
but this means exactly that A = limn→∞ An .
Proposition 1.23. The operations on F are extended from F by
continuity, i.e., if
an ), B = lim (bn ),
A = lim (
n→∞ n→∞
then
an + bn ),
A + B = lim ( an · bn ).
A · B = lim (
n→∞ n→∞
Proof. Exercise 19.
Exercises 17–19
Exercise 17. Prove that if {an } and {bn } are Cauchy sequences,
then so are
{an + bn }, {an − bn }, and {an bn }.
Exercise 18. Prove that if {an } ∼ {an } and {bn } ∼ {bn } are two
pairs of equivalent Cauchy sequences, then {an ± bn } ∼ {an ± bn } and
{an · bn } ∼ {an · bn }.
1.4. The field of p-adic numbers Qp

The basic example of a norm on the field Q of rational numbers is
the absolute value | · |. The induced metric d(x, y) = |x − y| is the
ordinary Euclidean distance on the number line, and as was explained
in §1.1, the field of real numbers R is the completion of Q with respect
to this norm.
Now let us ask ourselves the following question: Is the Euclidean
distance between rational numbers really the most “natural” one? Is
there any other way to describe the “closeness” between rationals? It
turns out that the answer to this question is YES.
The new ways of measuring distance between rational numbers

come from the following “arithmetical” construction.
Let p ∈ N be any prime number. Define a map | · |p on Q as
follows:

p− ordp x if x = 0,
(1.11) |x|p =
0 if x = 0,
where

the highest power of p which divides x, if x ∈ Z,
ordp x =
ordp a − ordp b, if x = a/b, a, b ∈ Z, b = 0
is the p-adic order of x (also called the p-adic valuation).
Remark 1.24. Notice that |·| can take only a “discrete” set of values,
namely, {pn , n ∈ Z} ∪ {0}.
Remark 1.25. If a, b ∈ N, then a ≡ b (mod pn ) if and only if
|a − b|p ≤ 1/pn .
Proposition 1.26. | · |p is a non-Archimedean norm on Q.
Proof. Property (1) in Definition 1.5 is obvious, and (2) follows from
(1.12) ordp (xy) = ordp (x) + ordp (y).
Let us verify (3). If x = 0 or y = 0, (3) is trivial, so assume x, y = 0.
Let x = a/b and y = c/d. Then we have
ad + bc
x+y = ,
bd
and
ordp (x + y) = ordp (ad + bc) − ordp (bd)
≥ min(ordp (ad), ordp (bc)) − ordp b − ordp d
= min(ordp a − ordp b, ordp c − ordp d)
= min(ordp x, ordp y).
Therefore,
|x + y|p = p− ordp (x+y) ≤ max(p− ordp x , p− ordp y )
= max(|x|p , |y|p ) ≤ |x|p + |y|p .
Observe that we have also proved that |·|p satisfies the strong triangle
inequality, i.e., it is non-Archimedean.
Remark 1.27. We shall see later that Q is not complete with respect
to the norm | · |p .
Remark 1.28. The norm | · |p1 is not equivalent to | · |p2 if p1 and p2
are different primes (indeed, for the sequence xn = (p1 /p2 )n we have
|xn |p1 → 0, but |xn |p2 → ∞).
We are now ready for the definition of the main protagonist of
these lectures. Let p be a fixed prime. We define Qp to be the
completion of Q with respect to the p-adic norm | · |p of (1.11). The
p-adic norm is extended to Qp according to Definition 1.20, and (Qp , |·
|p ) is a complete normed field. We call Qp the field of p-adic numbers.
The elements of Qp are equivalence classes of Cauchy sequences in Q
with respect to the extension of the p-adic norm. As has been pointed
out earlier, Q can be identified with the subfield of Qp consisting of
equivalence classes of constant Cauchy sequences.
For some a ∈ Qp let {an } be a Cauchy sequence of rational num-
bers representing a. Then by definition
(1.13) |a|p = lim |an |p .
n→∞
Therefore, the set of values that | · |p takes on Qp is the same as it

takes on Q, namely {pn , n ∈ Z}∪{0} — a phenomenon quite different
from what happens with the Euclidean absolute value which, when
extended from Q to R, takes all nonnegative real values. Moreover,
if |a|p = 0, then the sequence of norms {|an |p } must stabilize for
sufficiently large n!
Let us consider the series
d−m d−m+1
(1.14) m
+ m−1 + · · · + d0 + d1 p + b2 p2 + · · · ,
p p
where 0 < d−m < p and 0 ≤ di < p for all i > −m. Its partial sums
form a Cauchy sequence since for every > 0 we can choose N such
that p−N < , and for k > n > N we have
k
n k

di pi − di pi = di pi ≤ max (|di pi |p ) ≤ p−N < .
p p n<i≤k
−m −m n+1
Therefore, each series of the form (1.14) represents an element of Qp .

The converse statement is also true. We will show that each equiv-
alence class of Cauchy sequences in Q contains a unique canonical
representative Cauchy sequence — the sequence of partial sums of a

series in the form (1.14).
In order to describe its construction, we need the following lemma.
Lemma 1.29. If x ∈ Q and |x|p ≤ 1, then for any i there exists an
integer α ∈ Z such that |α − x|p ≤ p−i . The integer α can be chosen
in the set {0, 1, 2, . . . , pi − 1} and is unique if chosen in this range.
Proof. Let x = a/b, where a and b are relatively prime (this is de-
noted (a, b) = 1). Since |x|p ≤ 1, it follows that p does not divide b,
and hence b and pi are relatively prime. So, we can find integers m
and n such that mb + npi = 1. Let α = am. Then
|α − x|p = |am − a/b|p = |a/b|p |mb − 1|p
≤ |mb − 1|p = |npi |p = |n|p p−i ≤ p−i .
Finally, using the strong triangle inequality, we can add a multiple
of pi to the integer α to get an integer between 0 and pi for which
|α − x|p ≤ p−i still holds.
Theorem 1.30. Every equivalence class a in Qp satisfying |a|p ≤ 1
has exactly one representative Cauchy sequence {ai } such that
(1) ai ∈ Z, 0 ≤ ai < pi for i = 1, 2, . . .,
(2) ai ≡ ai+1 (mod pi ) for i = 1, 2, . . ..
Proof. Let {bi } be a Cauchy sequence representing a. We want to

find an equivalent sequence {ai } satisfying (1) and (2). Since
|bi |p −→ |a|p ≤ 1 as i → ∞,
throwing away several initial terms, if necessary, we may assume that
|bi |p ≤ 1 for all i.
For every j = 1, 2, . . . let N (j) be a positive integer such that
|bi − bi |p ≤ p−j , ∀i, i ≥ N (j).
Observe that we may take the sequence N (j) to be strictly increasing
with j, so N (j) ≥ j.
From Lemma 1.29, we can find integers aj , 0 ≤ aj < pj , such
that
1
|aj − bN (j) |p ≤ j .
p
Let us show that aj ≡ aj+1 (mod pj ) and (bi ) ∼ (aj ).

The first assertion follows since
|aj+1 − aj |p = |aj+1 − bN (j+1) + bN (j+1) − bN (j) − (aj − bN (j) )|p
≤ max(|aj+1 − bN (j+1) |p , |bN (j+1) − bN (j) |p , |aj − bN (j) |p )
≤ max(1/pj+1 , 1/pj , 1/pj ) = 1/pj ,
so that aj ≡ aj+1 (mod pj ).

To prove the second assertion, take any j; then for i ≥ N (j) we
have
|ai − bi |p = |ai − aj + aj − bN (j) − (bi − bN (j) )|p
≤ max(|ai − aj |p , |aj − bN (j) |p , |bi − bN (j) |p )
≤ max(1/pj , 1/pj , 1/pj ) = 1/pj .
Hence
|ai − bi |p −→ 0 as i → ∞.
Now, let us prove uniqueness. If {ai } is a different sequence

satisfying the requirements of the theorem with ai0 = ai0 for some i0 ,
then we have ai0 ≡ ai0 (mod pi0 ), since both ai0 and ai0 are between
0 and pi0 . Then it follows from (2) that for i > i0
ai ≡ ai0 ≡ ai0 ≡ ai (mod pi0 ),
i.e., ai ≡ ai (mod pi0 ). But this means exactly that

1
|ai − ai |p > , ∀i ≥ i0 ,
p i0
which implies that (ai ) ∼ (ai ).
If a ∈ Qp with |a|p ≤ 1, then it is convenient to write all the terms

ai of the representative sequence given by the previous theorem in the
following way:
ai = d0 + d1 p + . . . + di−1 pi−1 ,
where all the di ’s are integers in {0, 1, . . . , p − 1}. Our condition (2)
means precisely that
ai+1 = d0 + d1 p + . . . + di−1 pi−1 + di pi ,

where the “p-adic digits” d0 through di−1 are all the same as for
ai . Thus a is represented by the convergent (in the p-adic norm, of
course) series
∞
a= dn pn ,
n=0
which can be thought of as a number, written in the base p, that
extends infinitely far to the left or has infinitely many p-adic digits.
We will write
a = . . . dn . . . d2 d1 d0
and call this the canonical p-adic expansion or canonical form of a.
If |a|p > 1, then we can multiply a by a power of p (namely
by pm = |a|p ) so as to get a p-adic number a = apm that satisfies
|a |p = 1.
Then we can write
∞

(1.15) a= dn pn ,
n=−m
where d−m = 0 and bi ∈ {0, 1, 2, . . . , p − 1}, and represent the given

p-adic number a as a fraction in the base p with infinitely many p-adic
digits before the point and finitely many digits after:
(1.16) a = . . . dn . . . d2 d1 d0 . d−1 . . . d−m ;
this representation is called the canonical p-adic expansion of a.
The following proposition shows that the norm of a p-adic num-
ber is determined by the index of the first nonzero coefficient in its
canonical expansion.
∞
Proposition 1.31. If a = n=0 dn pn with dn = 0 for 0 ≤ n < k and
∞
dk = 0, then |a|p = p−k , and if a = n=−m dn pn , where d−m = 0,
then |a|p = pm .
Proof. By the definition (1.13), |a|p is a limit of the sequence of p-

adic norms of the partial sums of this series. In the first case we obtain
the constant sequence p−k , p−k , p−k , . . . since |dk |p = 1 (remember
that 0 < dk < p) and by the strong triangle inequality. Thus |a|p =

p−k . If a = ∞ n=−m dn p , where d−m = 0, the same argument shows
n
that |a|p = p .
m

Moreover, the notion of order defined in the beginning of §1.4 for

rational numbers can be extended to all p-adic numbers: for a ∈ Qp ,
ordp (a) is equal to the index of the first nonzero coefficient in the
canonical expansion of a.
Remark 1.32. The uniqueness assertion in Theorem 1.30 is some-

thing we do not have in the representation of numbers as infinite
fractions to the base g mentioned in §1.1, e.g. for g = 10
1.0000 . . . = 0.9999 . . . .
There are no such exceptions in the p-adic case. If two p-adic expan-
sions converge to the same p-adic number, they are the same, i.e., all
their digits are the same.
Definition 1.33. A p-adic number a ∈ Qp is said to be a p-adic

integer if its canonical expansion contains only nonnegative powers of
p.
The set of p-adic integers is denoted by Zp , so

∞

Zp = i
ai p .
i=0
It is easy to see (Exercise 21) that Zp = {a ∈ Qp | |a|p ≤ 1}.
Theorem 1.34. Every infinite sequence of p-adic integers has a con-

vergent subsequence.
Proof. Recall that a subsequence {xnk } of a sequence {xk } is given

by a sequence of positive integers {nk } such that n1 < n2 < n3 < . . ..
Let {xk } be a sequence in Zp . Let us write out the canonical
expansion of each term,
xk = . . . ak2 ak1 ak0 .
Since there are only finitely many possibilities for the digits ak0 (namely,
0, 1, . . . , p − 1), we can find b0 ∈ {0, 1, . . . , p − 1} and an infinite sub-
sequence of {xk }, denoted by {x0k }, such that the first digit of all
x0k is always b0 . The same trick yields b1 ∈ {0, 1, . . . , p − 1} and
a subsequence {x1k } of {x0k } for which the first two digits are b1 b0 .
This procedure can be continued, and we obtain b0 , b1 , b2 , . . . together

with a sequence of sequences
x00 ,x01 , x02 , . . . , x0s , . . . ,
x10 ,x11 , x12 , . . . , x1s , . . . ,
x20 ,x21 , x22 , . . . , x2s , . . . ,
...............
such that each sequence is a subsequence of the preceding one, and
such that each element of the (j + 1)th row begins with bj . . . b1 b0 .
For each j = 0, 1, . . . we have
xjj ∈ {xj−1 j , xj−1 j+1 , . . .}.
Therefore the diagonal sequence x00 , x11 , . . . is still a subsequence of
the original sequence, and it obviously converges to . . . b3 b2 b1 b0 .
Remark 1.35. It is not difficult to extend this result to bounded

sequences (see Exercise 25). The same result is true for bounded se-
quences of real numbers (Bolzano-Weierstrass Theorem), and a proof
for a sequence of real numbers 0 ≤ xn ≤ 1 can be modeled on the
above proof using the representation of real numbers by infinite dec-
imal fractions.
Exercises 20–25
Exercise 20. What is the cardinality of Zp ? Justify your answer.
Exercise 21. Prove that Zp = {a ∈ Qp | |a|p ≤ 1}.
Exercise 22. Find the p-adic norm and the p-adic expansion of
(1) 15, −1, −3 in Q5 ,
(2) 6! in Q3 ,
(3) 1/3! in Q3 .
Exercise 23. Find the p-adic expansion of 1/p. What about 1/pk ?
1.5. Arithmetical operations in Qp 27
Exercise 24. Find the p-adic expansion of 1/2 if p is an odd prime.
Exercise 25. Prove that any infinite bounded sequence in Qp has a

convergent subsequence.
1.5. Arithmetical operations in Qp

The p-adic expansion allows us to perform arithmetical operations in
Qp in a way very similar to that in R. Moreover, we will see that the
operations in Qp are, in fact, easier to perform than in R! Let
∞
∞

a= an p n , b= bn pn ,
n=−m n=−m
where an and bn are p-adic digits, a−m = 0, but possibly one or more
of the first digits b−m , b−m+1 , . . . are equal to 0. Then each
∞

a±b= (an ± bn )pn
n=−m
is a convergent series; however, in general it will not be in the canon-

ical form (1.15). The reduction to canonical form given by Theorem
1.30 corresponds to the standard addition (or subtraction) procedure
from right to left applied to p-adic numbers given in the form (1.16)
and uses a system of carries (similar to the one used for decimal frac-
tions).
In order to illustrate the addition algorithm, let us find the canon-
ical p-adic expansion of −1 in Qp . We have 1 = . . . 00001. Let
a = . . . a3 a2 a1 a0 satisfy 1 + a = 0 (then a = −1). Starting from
the right, we must have 1 + a0 = 0, but since a0 is in the range
{0, 1, . . . , p − 1}, the only way to achieve this is to have 1 + a0 = p
and to carry 1 to the left. Thus a0 = p−1. Continuing the procedure,
we see that all the an are equal to p − 1, i.e.,
−1 = . . . (p − 1)(p − 1)(p − 1).
Multiplication can be performed in a similar way. Let
∞
∞

a= an p n , b= bn pn
n=−m n=−k
be given in canonical form. Multiplying the series term by term and

rearranging terms, we obtain
∞
ab = un p n ,
n=−m−k
where
u−m−k = a−m b−k ,
u−m−k+1 = a−m+1 b−k + a−m b−k+1 ,
...............
This series again, in general, is not in the canonical form, but the
method of Theorem 1.30 allows us to reduce it to such a form. Again,
this corresponds to the standard multiplication procedure performed
on p-adic numbers given in the canonical form (1.16).
To illustrate division, suppose we have a, b ∈ Qp and b = 0.
Without loss of generality we may assume that b ∈ Zp , b = . . . b2 b1 b0
with b0 = 0, while
a = . . . a3 a2 a1 a0 . a−1 . . . a−k
is an arbitrary p-adic number. Since b0 = 0 and since the ring of
residues Z/pZ for a prime p is a field, we can always find a c−k ∈
{0, 1, . . . , p − 1} such that c−k b0 ≡ a−k (mod p). Continuing the
usual division procedure (carrying, if necessary, 1 to the left), we
obtain the quotient a/b in the canonical form.
It follows that if a = . . . a2 a1 a0 is a p-adic integer with a0 = 0,
then its multiplicative inverse a−1 is also a p-adic integer! (This
property of p-adic integers may seem strange at first glance, but it is
admittedly a nice one to have.) On the other hand, since
∞

p· ai pi = a0 p + a1 p2 + · · · = 1 + 0p + 0p2 + . . . ,
i=0
it follows that p has no multiplicative inverse in Zp (of course, p has a

multiplicative inverse in Qp (Exercise 23)!). A similar argument shows
that a p-adic integer whose first digit a0 is zero has no multiplicative
inverse in Zp . We summarize this in the following proposition.
Proposition 1.36. A p-adic integer
a = . . . a1 a 0 ∈ Z p
1.5. Arithmetical operations in Qp 29
has a multiplicative inverse in Zp if and only if a0 = 0.
We will denote the group of invertible elements in Zp by Z×

p,
∞

Z×
p = ai p | a0 = 0 .
i
i=1
This group is also called the group of p-adic units. By Exercise 26,
Z×
p = {x ∈ Zp | |x|p = 1}.
The following proposition follows at once from the definition of the

p-adic norm and Exercise 26.
Proposition 1.37. Let x be a p-adic number of norm p−n . Then x

can be written as the product x = pn u, where u ∈ Z×
p.
Notice that the arithmetical operations in Qp extend the ordinary

arithmetical operations on natural numbers (written in base p). The
familiar algorithms for addition, subtraction and multiplication are
simply pursued indefinitely from right to left. The p-adic division al-
gorithm also proceeds from right to left, and in that way it is different
from the familiar “long division” algorithm which begins by finding
the digit in the highest position and proceeds from left to right.
Here are some examples of arithmetical operations in Q7 :
. . . 615
. . . 46530.25 . . . 263 . . . 153 ). . . 421
+
. . . 20656.41 × . . . 161
. . . 154
. . . 00516.66 . . . 445 . . . 230
. . . 46530.2 . . . 141 . . . 153
− . . . 263 . . . 400
. . . 20656.4
. . . 25540.5 ... 455 ...4
......
Exercises 26–31
Exercise 26. Prove that Z×

p = {x ∈ Zp | |x|p = 1}.
Exercise 27. If a ∈ Qp has the canonical p-adic expansion
. . . an . . . a2 a1 a0 . a−1 . . . a−m ,
what is the canonical p-adic expansion of −a?
Exercise 28. The integers 2, 3, 4 are invertible in Z5 . Find the 5-

adic expansions of their inverses. Find the expansion of 1/3 in Z7 .
Exercise 29. Find the canonical p-adic expansion of:

(1) . . . 1246 × . . . 6003 in Q7 to 4 digits,
(2) 1 : . . . 1323 in Q5 to 4 digits,
(3) 900 − . . . 312.3 in Q11 to 4 digits.
Exercise 30. Find the p-adic norm of (pn )!.
Exercise 31.* Find the p-adic norm of n!.
1.6. The p-adic expansion of rational numbers

Any rational integer is also a p-adic integer (simply write its expan-
sion in base p). However, there are p-adic integers among rational
fractions! We have seen that
∞

−1 = (p − 1) pi ,
i=0
so that
∞
1 1
pi = , = . . . 1111,
i=0
1−p 1−p
which is in Zp . Note that the p-adic expansion of this p-adic integer
is infinite! See Exercise 34 for necessary and sufficient conditions for
a p-adic expansion to terminate.
1.6. The p-adic expansion of rational numbers 31
The following theorem shows that we can recognize rational num-

bers by their p-adic expansion just as we recognize rationals among
reals by their decimal expansion.
Theorem 1.38. The canonical p-adic expansion (1.16) represents a
rational number if and only if it is eventually periodic to the left.
Proof. Assume that the canonical p-adic expansion is eventually pe-

riodic. Multiplying (if necessary) the given p-adic number x by a
power of p and subtracting a rational number, we may consider the
case in which x ∈ Zp has a periodic expansion of the form
x = x0 + x1 p + x2 p2 + · · · + xk−1 pk−1 + x0 pk + x1 pk+1 . . . .
The number a = x0 + x1 p + x2 p2 + · · · + xk−1 pk−1 is a rational integer
and x can be expressed in the form
1
x = a(1 + pk + p2k + . . . ) = a ,
1 − pk
and hence x is a rational number.
Conversely, suppose that
a
(1.17) = xi p i ∈ Z p .
b
i≥0
We may assume that a and b are relatively prime integers and b is

relatively prime to p. Since (b, pn ) = 1, there exist cn , dn such that
1 = cn b + dn pn .
Multiplying both sides by a, we obtain
a = acn b + adn pn .
By adding a multiple of pn to the integer acn , we obtain two integers,
An with 0 ≤ An ≤ pn − 1 and rn such that the equality
a = A n b + rn p n
holds. Dividing both sides by b, we obtain
a rn
= An + pn .
b b
Hence rn = (a − An b)/pn , and therefore
a − (pn − 1)b a
n
≤ rn ≤ n .
p p
For sufficiently large n, this implies −b ≤ rn ≤ 0, which means that

rn takes only finitely many values. Now we can write
a rn rn+1
(1.18) = An + pn = An+1 + pn+1 .
b b b
Since An+1 − An = pn ( rn −pr b

n+1
) is an integer and (b, pn ) = 1, the
rn −prn+1
expression b is an integer. Hence An+1 ≡ An (mod pn ), and
by the uniqueness of Theorem 1.30 the sequence {An } is the sequence
of partial sums of the canonical p-adic representation (1.17) of a/b.
Thus An+1 = An + xn pn , and (1.18) implies rn = xn b + prn+1 for all
n. Since rn takes only finitely many values, there exist an index m
and a positive integer P such that rm = rm+P ; hence
(1.19) xm b + prm+1 = xm+P b + prm+P +1 ,
so that
(xm − xm+P )b = p(rm+P +1 − rm+1 ).
Since (b, p) = 1, it follows that p divides xm − xm+P . But both

xm and xm+P are digits in {0, 1, . . . , p − 1}; therefore xm = xm+P .
If we substitute this into (1.19), we also see that rm+1 = rm+P +1 .
Repeating this argument, we obtain
rn = rn+P and xn = xn+P (n ≥ m),
which proves that not only the sequence of digits xn , but also the
sequence of numerators rn , has a period of length P for n ≥ m.
Is it possible to determine from a p-adic expansion of a rational

number whether it is positive or negative? The answer is YES, and
it is given in Exercise 33.
1.7. Hensel’s Lemma and congruences 33
Exercises 32–34
Exercise 32. Prove that

(1) Zp ∩ Q = {a/b ∈ Q : p b},
(2) Z×
p ∩ Q = {a/b ∈ Q : p ab}.
Exercise 33.* Let r ∈ Q. Prove that for an appropriate k ≥ 1, rpk

is a p-adic integer and its p-adic expansion can be represented in the
form . . . aaaaab, where the fragments a and b have the same number
of digits. Prove that r > 0 is equivalent to b > a in the usual sense
(as integers written in base p).
Exercise 34. Prove that the p-adic expansion of a ∈ Qp terminates

(i.e., ai = 0 for all i greater than some N ) if and only if a is a non-
negative rational number whose denominator is a power of p.
1.7. Hensel’s Lemma and congruences

√
Let us extract 6 in Q5 ; this means we want to find a sequence of
5-adic digits a0 , a1 , a2 , . . . , 0 ≤ ai ≤ 4, such that
(1.20) (a0 + a1 · 5 + a2 · 52 + . . .)2 = 1 + 1 · 5.
From (1.20) we obtain a20 ≡ 1 (mod 5), which implies a0 = 1 or

4. If a0 = 1, then
2a1 · 5 ≡ 1 · 5(mod 52 ) ⇒ 2a1 ≡ 1 (mod 5) ⇒ a1 = 3.
At the next step we have
1 + 1 · 5 ≡ (1 + 3 · 5 + a2 · 52 )2 ≡ 1 + 1 · 5 + 2a2 · 52 (mod 53 ) ,
which implies 2a2 ≡ 0 (mod 5) and therefore a2 = 0.
So, we get a series
a = 1 + 3 · 5 + 0 · 52 + . . . ,
where each ai after a0 is uniquely determined.
If we choose a0 = 4, then we obtain the solution
−a = 4 + 1 · 5 + 4 · 52 + 0 · 53 + . . . .
It is not very difficult to see that there exist numbers in Q5 which

have no square root (for example 2 + 1 · 5).
The above method of solving equations (like x2 −6 = 0 in Q5 ) can
be generalized by using an extremely important result called “Hensel’s
Lemma”.
Generalizing Remark 1.25, we say that a and b ∈ Qp are congruent
modulo pn and write
a ≡ b (mod pn )
if and only if |a − b|p ≤ 1/pn .
Theorem 1.39. (Hensel’s Lemma) Let F (x) = c0 + c1 x + . . . + cn xn

be a polynomial whose coefficients are p-adic integers. Let
F (x) = c1 + 2c2 x + 3c3 x2 + . . . + ncn xn−1
be the derivative of F (x). Suppose a0 is a p-adic integer which satis-
fies F (a0 ) ≡ 0 (mod p) and F (a0 ) ≡ 0 (mod p). Then there exists a
unique p-adic integer a such that F (a) = 0 and a ≡ a0 (mod p).
Proof. We will prove the existence of a by constructing its canonical

p-adic expansion a = b0 + b1 p + b2 p2 + . . . inductively. At the kth step
of induction, we will find ak = b0 + · · · + bk pk , the kth approximation
of a, by using a p-adic version of Newton’s method (cf. the remark at
the end of this proof). Each ak will not be a true root of F (x), but
only a “root modulo pk+1 ” (i.e., we will have F (ak ) ≡ 0 (mod pk+1 ))
for all k. In the limit as k → ∞ we will obtain a, the required true
root of F .
More precisely, we will prove the following statement by induction
on k:
S(k): there exists a p-adic integer of the form
ak = b0 + b1 p + · · · + bk pk
(whose digits bi are in {0, 1, . . . , p − 1} for all i) such that
F (ak ) ≡ 0 (mod pk+1 ) and ak ≡ a0 (mod p).
The base of induction is obvious: taking b0 equal to the first

p-adic digit of a0 , we will have a0 ≡ a0 and F (a0 ) ≡ 0 (mod p).
Now let us perform the induction step, i.e., prove that S(k − 1)
implies S(k). To do this, we set ak = ak−1 + bk pk for some (as yet
unknown) digit bk satisfying 0 ≤ bk < p and expand F (ak ), ignoring
terms divisible by pk+1 :

n
F (ak ) = F (ak−1 + bk pk ) = ci (ak−1 + bk pk )i
i=0

n
= c0 + ci (aik−1 + iai−1 k
k−1 bk p + terms divisible by p
k+1
)
i=1
≡ F (ak−1 ) + bk pk F (ak−1 ) (mod pk+1 ).
Since F (ak−1 ) ≡ 0 (mod pk ) by the inductive assumption, we can

write
F (ak ) ≡ αk pk + bk pk F (ak−1 ) (mod pk+1 )
for some integer αk ∈ {0, 1, . . . , p − 1}. Thus we come to the following
equation for the unknown digit bk :
αk + bk F (ak−1 ) ≡ 0 (mod p),
which we can easily solve provided F (ak−1 ) ≡ 0 (mod p). But this
is indeed the case because we obviously have ak−1 ≡ a0 (mod p), so
that
F (ak−1 ) ≡ F (a0 ) ≡ 0 (mod p).
Dividing by F (ak−1 ), we can find the required digit bk ,
−αk
bk = (mod p),
F (ak−1 )
for which we will have F (ak ) ≡ 0 (mod pk+1 ), completing the induc-
tion step.
Now, let
a = b0 + b1 p + b2 p2 + . . . .
Observe that F (a) = 0 since for all k we have
F (a) ≡ F (ak ) ≡ 0 (mod pk+1 ).
The uniqueness of a follows from the uniqueness of the sequence {ak }.

Remark 1.40. The second condition (F (a0 ) ≡ 0 (mod p)) in The-

orem 1.39 is essential (see Exercise 35). However, there are many
different versions of Theorem 1.39 in the literature, all of which are
referred to as “Hensel’s Lemma”. Exercise 40 gives a version that can
be used when the hypothesis F (a0 ) ≡ 0 (mod p) does not hold.
Remark 1.41. The method of approximation used in the proof of
Helsel’s Lemma essentially coincides with Newton’s method of find-
ing a real root of a polynomial f (x) with real coefficients. In the
real case, if f (an−1 ) = 0, according to Newton’s method, the next
approximation an is given by the formula
f (an−1 )
an = an−1 − .
f (an−1 )
The correction term looks very much like the “correction term” in the
proof of Hensel’s Lemma:
αn pn F (an−1 )
bn pn ≡ − ≡− (mod pn+1 ).
F (an−1 ) F (an−1 )
In one respect, however, Hensel’s Lemma is better than Newton’s
method in the real case: in the p-adic case the convergence to a root
of the polynomial is guaranteed by universal conditions on the approx-
imate solution a0 whose form does not depend on the polynomial. In
the real case, Newton’s method converges if the approximate solution
is sufficiently close to the actual root, but the condition of closeness
√ = x − x√and the
3
depends on the polynomial.√For example, for f (x)
unfortunate choice a0 = 1/ 5, we get a1 = −1/ 5, a2 = 1/ 5, etc.,
so the sequence {an } does not converge.
Let us recall that in Theorem 1.30 the canonical expansion of p-

adic numbers came out of a sequence of congruences. Hensel’s Lemma
confirms this connection. The following theorem makes the connec-
tion between p-adic numbers and congruences even more prominent.
Theorem 1.42. A polynomial with integer coefficients has a root in
Zp if and only if it has an integer root modulo pk for any k ≥ 1.
Proof. Let F (x) be a polynomial with coefficients in Z. Suppose

a ∈ Zp is its root, i.e.,
(1.21) F (a) = 0.
By Theorem 1.30 there exists a sequence of integers {a1 , a2 , . . . , ak , . . .},

where ak = b0 + b1 p + b2 p2 + · · · + bk−1 pk−1 , such that
a ≡ ak (mod pk ).
Then F (ak ) ≡ F (a) (mod pk ) and F (a) = 0 imply
(1.22) F (ak ) ≡ 0 (mod pk ).
Conversely, suppose the congruence (1.22) has an integer solution

ak for any k ≥ 1. According to Theorem 1.34, the sequence {ak }
contains a convergent subsequence {aki }, limi→∞ aki = a. We want
to show that a is a solution of equation (1.21). Since a polynomial is
a continuous function, we have
F (a) = lim F (aki )

i→∞
(here we just use the fact that the limit of the sum is the sum of
the limits and the limit of the product is the product of limits, i.e.,
Proposition 1.23). On the other hand,
F (aki ) ≡ 0 (mod pki ).
Therefore limi→∞ F (aki ) = 0, and thus F (a) = 0.
A practical consequence of Theorem 1.42 is the following. If a

polynomial with integer coefficients has no roots modulo p, then it
has no roots in Zp . It is usually not too hard to find its roots modulo
p if it has any. If a root modulo p is not a root of the derivative
modulo p, then by Hensel’s Lemma, we can find a root in Zp .
We say that a rational integer a not divisible by p is a quadratic
residue modulo p if the congruence
x2 ≡ a (mod p)
has a solution in {1, 2, . . . , p − 1}. Otherwise a is called a quadratic

nonresidue.
Proposition 1.43. A rational integer a not divisible by p has a

square root in Zp (p = 2) if and only if a is a quadratic residue
modulo p.
Proof. Let P (x) = x2 − a. Then P (x) = 2x. If a is a quadratic

residue, then
a ≡ a20 (mod p)
for some a0 ∈ {1, 2, . . . , p − 1}. Hence P (a0 ) ≡ 0 (mod p). But
P (a0 ) = 2a0 ≡ 0 (mod p)
automatically since (a0 , p) = 1, so that the solution in Zp exists by
Hensel’s Lemma. Conversely, if a is a quadratic nonresidue, by The-
orem 1.42 it has no square root in Zp .
√
For example −1 is in Z5 since −1 = 4 − 5 ≡ 22 (mod 5) is a
√
quadratic residue modulo 5, while −1 is not in Z3 since −1 = 2 − 3
√
is a quadratic nonresidue modulo 3. Is p in Zp ?
Exercises 35–44
Exercise 35. Construct a polynomial with integer coefficients which

has a root modulo 2 but no roots in Q2 .
Exercise 36. Prove that if p = 2, a p-adic unit

u = c0 + c1 p + c2 p 2 + . . .
is a square in Zp if and only if c0 is a quadratic residue modulo p.
Exercise 37. Let p = 2 be a prime. Denote (Q× ×

p ) = {a | a ∈ Qp }.
2 2
× × 2
Prove that the quotient group Qp /(Qp ) has order 4, and find a
complete set of coset representatives for it.
Exercise 38. Prove that the equation x3 − 1 = 0 has a solution

a = 1 in Z7 and find the first 3 digits in its canonical expansion.
Exercise 39. Prove that the equation x5 − 1 = 0 has no solution

a = 1 in Q7 . You must explain why such roots of unity, if they existed,
must have been in Z7 , not merely in Q7 !
Exercise 40. Let F (x) be a polynomial whose coefficients are in Zp ,

and suppose there exists a a0 ∈ Zp such that |F (a0 )|p ≤ |F (a0 )|2p .
1.8. Algebraic properties of p-adic integers 39
Then there exists a unique a ∈ Zp such that F (a) = 0 and |a − a0 |p ≤

| FF(a0)
(a0 ) |p .
Exercise 41. Give an example of when the original version of Hensel’s

Lemma (Theorem 1.39) cannot be used to find a root of a polynomial
equation, but Exercise 40 can be used and gives a root.
Exercise 42. Show that a unit u ∈ Z2 is a square if and only if

u ≡ 1 (mod 8).
√
Exercise 43. Prove that Q2 contains −7 and Qp (p > 2) contains
√
1 − p.
Exercise 44. Prove that the quotient group Q× × 2

2 /(Q2 ) has order 8,
is isomorphic to Z/2Z ⊕ Z/2Z ⊕ Z/2Z and admits a complete set of
representatives {±1, ±5, ±2, ±10}.
1.8. Algebraic properties of p-adic integers

We have already seen that the p-adic integers Zp differ in many ways
from the ordinary integers Z. Here we will see that their algebraic
properties are just as good, if not better, than those of Z. The ordi-
nary integers Z form a commutative ring (for the definition see §1.2).
Recall that a nonempty subset I of a ring R is said to be an ideal
if I is a subgroup of R under addition and for any x ∈ I and r ∈ R,
r · x ∈ I. For example, the set mZ of all integers divisible by a given
number m forms an ideal in the ring Z.
An ideal is called maximal if it is not contained in any other
proper ideal. In the example above, the ideal mZ is maximal if and
only if m is a prime number.
For a ring R and an ideal I ⊂ R, one can define the factor R/I
as the set of additive cosets with canonically defined addition and
multiplication. If R is a commutative ring with the multiplicative
identity, R/I is a field if an only if I is a maximal ideal. For example,
Z/pZ is the field of residues modulo p, which is the only field with p
elements. For details see [6, §3.5].
A commutative ring without zero divisors is called an integral
domain.
Proposition 1.44. The ring Zp is an integral domain.
Proof. This follows because Zp is contained in Qp , which is a field,

and hence contains no zero divisors.
Let Fp = Z/pZ be the finite field with p elements. The map

∞

a= ai pi → a0
i=0
defines a ring homomorphism Zp → Fp called reduction modulo p.

This homomorphism is surjective, and its kernel is
∞ ∞

{a ∈ Zp | a0 = 0} = i
ai p = p i
ai+1 p = pZp .
i=1 i=0
Since the quotient is a field, the kernel pZp is a maximal ideal of the
ring Zp .
Corollary 1.45. The ring Zp has a unique maximal ideal, namely,
pZp = Zp \ Z×
p.
Proof. Suppose I is another maximal ideal. Since pZp is maximal,

I must contain an element from its complement, a ∈ Z×
p . Since I is
an ideal, 1 = a · a−1 ∈ I, but then I = Zp .
Proposition 1.46. The ring Zp is a principal ideal domain. More
precisely, its ideals are the principal ideals {0} and pk Zp for all k ∈ N.
Proof. Let I = {0} be an ideal in Zp , and let 0 = a ∈ I be an element

of maximal norm (since the norm takes discrete set of values, such
an element can always be found). Assume that |a|p = p−k for some
k ∈ N. Then a = pk , where is a unit. Then pk = −1 a ⊂ I, and
hence (pk ) = pk Zp ⊂ I. Conversely, for any b ∈ I, |b|p = p−w ≤ p−k .
We can write
b = pw = pk pw−k ∈ pk Zp .
Therefore, I ⊂ pk Zp , and hence I = pk Zp .
We discussed the existence of square roots in Qp in §1.7 as an

application of Hensel’s Lemma. Another application has to do with
finding which roots of unity are in Qp .
1.8. Algebraic properties of p-adic integers 41
Recall that an element in the field ζ is called an mth root of unity

if ζ m = 1; it is called a primitive mth root of unity if, in addition,
ζ n = 1 for 0 < n < m.
Proposition 1.47. For any prime p and any positive integer m rel-
atively prime to p, there exists a primitive mth root of unity in Qp if
and only if m|(p − 1). In the latter case, every mth root of unity is
also a (p − 1)th root of unity. The set of (p − 1)th roots of unity is a
cyclic subgroup of Z×p of order (p − 1).
Proof. Let m|(p − 1); then p − 1 = km for k ≥ 1, and therefore any

mth root of 1 is also a (p − 1)th root of 1. Let
f (x) = xp−1 − 1, f (x) = (p − 1)xp−2 .
Take x0 ∈ Z×
p to be any rational integer satisfying 1 ≤ x0 ≤ p − 1.
Then
f (x0 ) ≡ 0 (mod p) and f (x0 ) ≡ 0 (mod p)
since |f (x0 )|p = 1, and Hensel’s Lemma applies, giving exactly p − 1
solutions, which are (p − 1)th roots of 1. The first digits of these
roots are 1, 2, . . . , p − 1. Conversely, if α ∈ Qp is an mth root of 1,
αm = 1, we must have |α|p = 1, i.e., α ∈ Zp . If α0 is its first digit,
then α0m ≡ 1 (mod p); hence m divides p − 1, the order of (Z/pZ)× .
Since a polynomial with coefficients in a field can only have as many
roots as its degree ([6, Lemma 5.3.2]), the polynomial xp−1 −1 cannot
have more than p − 1 roots, and these roots must be all the roots of
unity in Qp . It is clear that the roots of unity form a group under
multiplication. Finally, since any finite subgroup of the multiplicative
group of any field is cyclic ([6, Lemma 7.1.6]), the group of (p − 1)th
roots of unity is a cyclic subgroup of Z× p of order (p − 1).
The (pn )th roots of unity cannot be handled by means of Hensel’s

Lemma (why?) and will be discussed in Theorem 3.36.
The (p − 1)th roots of unity are related to the signum function
sgnp (x) introduced in the next theorem.
n
Theorem 1.48. For any x ∈ Zp the limit limn→∞ xp exists. This
limit is denoted by sgnp (x) and has the following properties:
(a) sgnp (x) depends only on the first digit in the canonical p-
adic expansion of x, x0 ;
(b) sgnp (xy) = sgnp (x) · sgnp (y);
(c) sgnp (x) = 0 if x0 = 0, and it is a (p − 1)th root of 1 if
x0 = 0.
Proof. Let x0 ∈ {1, 2, . . . , p − 1}. First we show that the sequence

n
{xp0 } converges. By Euler’s Theorem,
ϕ(pn )
x0 ≡ 1 (mod pn ),
where ϕ is Euler’s ϕ-function: for a positive integer m, ϕ(m) is equal
to the number of integers smaller than m and relatively prime to m.
Observe that since p is a prime, we have ϕ(pn ) = pn − pn−1 . Thus,
n
−pn−1 n n−1
xp0 ≡ 1 (mod pn ), xp0 ≡ xp0 (mod pn ),
and hence n
p pn−1 1
x0 − x0 ≤ n .
p p
n
Since 1/pn → 0 as n → ∞, the sequence {xp0 } is Cauchy, and by the
completeness of Zp , it converges to a limit in Zp , which we denote by
n
sgnp (x0 ) = lim xp0 .
n→∞
The limit obviously exists for x0 = 0, so sgnp (x) is defined for x0 ∈

{0, 1, 2, . . . , p − 1}, and sgnp (0) = 0. Next we show that the limit
exists for all x ∈ Zp and is defined by the first digit x0 of x. For this
we will need the following lemma.
Lemma 1.49. Suppose x ∈ Zp with the first digit x0 . Then we have
|xp − xp0 |p ≤ p−1 |x − x0 |p .
Proof of the lemma. Let x = x0 + α, with |α|p ≤ p−1 . Then

p p−1 p p−2 2 p p
xp − xp0 = x0 α + x0 α + · · · + α
1 2 p

p p−1 p p−2 p p−1
= (x − x0 ) x + x α + ···+ α .
1 0 2 0 p

Since | pj xp−j
0 αj−1 |p ≤ p−1 for j ≥ 1, by the strong triangle inequal-
ity we obtain |xp − xp0 |p ≤ p−1 |x − x0 |p .
1.9. Ostrowski’s Theorem 43
Applying the lemma, we obtain

n n−1
p pn n−1
x − x0 ≤ p−1 xp − xp0 ≤ · · · ≤ p−n |x − x0 |p ,
p p
n
pn
which implies that limn→∞ x exists and is equal to limn→∞ xp0 .
Thus we have defined sgnp (x) for all x ∈ Zp , and property (a) of
Theorem 1.48 is satisfied. Property (b) follows from the property of
limits:
n n n n n
lim (xy)p = lim (xp )(y p ) = lim xp lim y p .
n→∞ n→∞ n→∞ n→∞
It remains to show that if x0 ∈ {1, 2, . . . , p − 1}, then sgnp (x0 ) is a

(p − 1)th root of 1. Using property (b) and Fermat’s Little Theorem
(which is Euler’s Theorem for n = p), we obtain
sgnp−1
p (x0 ) = sgnp (xp−1
0 ) = sgnp (1) = 1.
The values of sgnp (x) are thus solutions of the equation y p − y = 0.
Since Qp is a field, this equation cannot have more than p solutions in
Qp , and hence in Zp . Consequently, the only solutions of this equation
are the values of the signum function.
1.9. Metrics and norms on the rational

numbers. Ostrowski’s Theorem
We have seen that the field Q admits the p-adic norm | · |p for each
prime p, as well as the ordinary absolute value |·| (which is sometimes
denoted by | · |∞ for p = ∞, also referred to as the infinite prime).
We shall prove now that there are no other norms on Q, and hence
the only completions of Q are Qp for all prime p, and R = Q∞ .
Theorem 1.50. (Ostrowski’s Theorem). Every nontrivial norm ·

on Q is equivalent to | · |p for some prime p or for p = ∞.
Proof. Suppose first that · is Archimedean, i.e., there exists a

positive integer n such that n > 1, and let n0 be the least such n.
Then we can write n0 = nα 0 for some positive real number α.
Now, write any positive integer n to the base n0 , i.e., in the form
n = a0 + a1 n0 + a2 n20 + . . . + as ns0 ,
where 0 ≤ ai < n0 , i = 0, . . . s, and as = 0. Then

n ≤a0 + a1 n0 + a2 n20 + . . . + as ns0
=a0 + a1 nα
0 + a2 n0 + . . . + as n0 .
2α sα
Since all of the digits ai are less than n0 , by our choice of n0 , we have
ai ≤ 1, and hence
n ≤ 1 + nα 2α sα
0 + n0 + . . . + n0
−α −2α
≤ nsα
0 (1 + n0 + n0 + . . . + n−sα
0 )
∞
1 i
≤ nα ,
i=0
nα
0
because n ≥ ns0 . The expression in brackets is a finite positive con-

stant independent of n, which we call C. Thus
n ≤ Cnα for all n = 1, 2, . . . .
The same argument with nN in place of n yields
√
nN ≤ CnN α ⇒ n ≤ Cnα .
N
Letting N → ∞ for n fixed, we obtain

(1.23) n ≤ nα .
Let us prove the opposite inequality. First, observe that
ns+1
0 > n ≥ ns0 .
Since
(s+1)α
n0 = ns+1
0 = n + n0
s+1
− n ≤ n + ns+1
0 − n,
we have
(s+1)α
n ≥ ns+1
0 − n0
s+1
− n ≥ n0 − (ns+1
0 − n)α ,
because
ns+1
0 − n ≤ (ns+1
0 − n)α
as was proved in (1.23). Thus
(s+1)α
n ≥ n0 − (ns+1
0 − ns0 )α (since n ≥ ns0 )
1 α
) ] = C n0 ≥ C nα
(s+1)α (s+1)α
= n0 [1 − (1 −
n0
for some positive constant C that does not depend on n.
1.9. Ostrowski’s Theorem 45
As before, we now use this inequality for nN , take N th roots, and

let N → ∞, obtaining
(1.24) n ≥ nα .
From (1.23) and (1.24), we deduce that n = nα for all n ∈ N.

Using property (2) of the norm, we readily see that x = |x|α for
all x ∈ Q. In view of Proposition 1.10, we can conclude that such a
norm is equivalent to the absolute value | · |.
Now suppose that · is non-Archimedean, i.e., we have n ≤ 1

for all positive integers n. Because we have assumed that · is
nontrivial, we can find n0 , the least n such that n < 1. Observe that
n0 must be a prime number, because if n0 = n1 n2 , with n1 , n2 < n0 ,
then n1 = n2 = 1, and so n0 = n1 n2 = 1. Denote the
prime number n0 by p.
Next, we will prove that if n is not divisible by p, then n = 1.
Write n = rp + s with 0 < s < p. By the minimality of p, s = 1.
We also have rp < 1 since p < 1 (by choice) and r ≤ 1 (by the
non-Archimedean property, since r is an integer). Consequently,
n − s < s,
and by Proposition 1.15, n = s = 1. Finally, given any n ∈ Z,
we can write n = pv n , where p does not divide n . Hence
n = pv n = pv .
Let ρ = p < 1. Then ρ = (1/p)α for some positive real α. Therefore
n = |n|α
p.
Now, it is easy to show (using property (2) of the norm) that the
same formula holds with any nonzero rational number x in place of
n. In view of Proposition 1.10, we have · ∼ | · |p and this concludes
the proof of the theorem.
Proposition 1.51. (Product Formula). Let Q× = Q \ {0}. For any
x ∈ Q× we have
|x|p = 1,
p≤∞
where the product is taken over all primes of Q including the “prime
at infinity”.
Proof. It is sufficient to prove this formula when x is a positive

integer; the rest follows from the multiplicative property of the norm.
So, suppose that x = pa1 1 · pa2 2 · · · pakk . Then |x|q = 1 if q = pi ,
|x|pi = p−a
i
i
for i = 1, . . . , k, and |x| = pa1 1 · pa2 2 · · · pakk . The result
follows.
The product formula establishes a close relationship between the

norms on Q. For instance, if we know the values of all but one norm,
this allows us to recover the value of the missing one. This is very
important in many applications to algebraic geometry.
Suppose we want to find a root of a polynomial in Q. Evidently,
if there are roots in Q, then there are roots in R and in all Qp . Hence
we can certainly conclude that there are no rational roots if there is
some p ≤ ∞ for which there are no p-adic roots (again, “∞-adic”
means “real”). A converse statement would be more interesting, but
is it true? If a polynomial has p-adic roots for all p including ∞, does
it follow that it has a rational root? Here is a simple example when
such a converse statement holds.
Proposition 1.52. A number x ∈ Q is a square if and only if it is a

square in every Qp , p ≤ ∞.
Proof. For any x ∈ Q× we have

x=± pordp (x) .
p<∞
Notice that x is a square in R if and only if it is positive. In Qp we

can write x = pordp (x) u, where u ∈ Z×p (Proposition 1.37). Then x is
a square in Qp if and only if ordp (x) is even and u = v 2 for some unit
v ∈ Z×p . If we write out the factorization, we see that x is a square
in Q if and only if it is a square in each Qp .
This is a manifestation of the so-called Local-to-Global Princi-

ple, which asserts that the existence or nonexistence of solutions in
Q (global solutions) of a Diophantine equation can be detected by
studying, for each p ≤ ∞, the solutions in Qp (local solutions). Un-
fortunately, this principle is not universal, but it holds in some impor-
tant cases, for instance for quadratic forms in several variables (the
1.10. A digression: what about Qg if g is not a prime? 47
Hasse-Minkowski Theorem, [17]). Exercise 46 gives an example when

this principle does not hold.
Exercises 45–46
Exercise 45. Two fields F and K are called isomorphic if there

exists a bijective map ϕ : F → K such that
ϕ(a + b) = ϕ(a) + ϕ(b), ϕ(a · b) = ϕ(a) · ϕ(b).
(1) Prove that Qp and R are not isomorphic.
(2)* Prove that if p = q are two primes, then Qp and Qq are not
isomorphic.
Exercise 46. Prove that an equation

(x2 − 2)(x2 − 17)(x2 − 34) = 0
has a root in Qp for all r ≤ ∞ but not in Q.
1.10. A digression: what about Qg if g is not a

prime?
In order to fully appreciate the beauty of p-adic numbers, let us see
what happens if we use a nonprime number g instead of a prime p.
In order to use the formula (1.11) to define | · |g , we have to be a
little more careful in the definition of ordg (x) for a rational number
a. If x ∈ Z, ordg (x) is still the highest power of g which divides x. If
x = a/b, the definition is different from the case in which g = p is a
prime, and in order to give it, we need the following lemma.
Lemma 1.53. Let x = a/b, a, b ∈ N, (a, b) = 1. Then there exist

a unique integer v and a pair of integers a and b such that a/b =
g v a /b , g a (a , b ) = (g, b ) = 1.
Proof. Obviously, if g a and (g, b) = 1, then we can choose v = 0.

If g|a, denote by g φ , where φ ≥ 1, the highest power of g that
divides a. Putting a = ag −φ and b = b, we obtain g a and
(a , b ) = (g, b ) = 1, and furthermore

a a
= gφ ,
b b
which proves the assertion with v = φ > 0.
Now assume that g a and (g, b) > 1. Then b can be factored
as b = b1 b2 , (b1 , b2 > 0) so that all prime factors of b1 divide g and
(b2 , g) = 1. Similarly, g can be written as g = g1 g2 (g1 , g2 > 0) so
that all prime factors of g1 divide b1 , but (g2 , b1 ) = (g2 , b) = 1. There
is a smallest positive integer ψ such that b1 |g ψ and hence also b1 |g1ψ .
In the equation
a gψ gψ
g ψ = 1 2 a,
b b1 b2
ψ
the quotient g1 /b1 is an integer. Therefore if we put
g1ψ ψ
a = g a and b = b2 ,
b1 2
we obtain
a a
= g −ψ , g a , (a , b ) = (g, b ) = 1,
b b
proving the assertion with v = −ψ < 0.
Now for x = a/b, we define ordg (x) to be the integer v from

Lemma 1.53, and we define the corresponding norm as
(1.25) |a/b|g = g −v .
This norm, however, will not satisfy the multiplicative property (2)
of Definition 1.5. For example,

1
= 102 , 1 = 102 , but 1 · 1 = 1 = 103 ;
20 50 20 50 1000
10 10 10 10
therefore
1 1
· < 1 · 1 .
20 50 20 50
10 10 10
In general
(1.26) |ab|g ≤ |a|g |b|g ,
and | · |g is not a norm but a so-called pseudo-norm (see Exercise 47).
Nevertheless, d(x, y) = |x − y|g is still a metric, and one can consider
the completion of Q with respect to this metric. Denoted by Qg , it is
1.10. A digression: what about Qg if g is not a prime? 49
a ring but not a field if g is not a prime (see Exercise 48). Of course,
if g = p is a prime, then the definition in (1.25) coincides with the
definition in (1.11).
The following theorem is due to Hensel:
Theorem 1.54. If g = p1 p2 . . . pk is a product of distinct primes,
then Qg = Qp1 ⊕ · · · ⊕ Qpk , the direct sum of p-adic fields.
Proof. We will construct this isomorphism in the case g = 10, p1 =

2, p2 = 5, but the general case is handled similarly without any
complications.
Consider a Cauchy sequence in Q relative to | · |10 . It defines a
10-adic number
(10)
A = lim an ,
n→∞
and the existence of the 10-adic limit implies the existence of the
2-adic and the 5-adic limits which we denote by
(2) (5)
A2 = lim an , A5 = lim an ,
n→∞ n→∞
respectively. Conversely, the existence of the limits A2 and A5 evi-
dently implies that of A. It is easy to see that the digits of A2 and
A5 do not depend on the Cauchy sequence {an } by means of which
A was defined.
In particular, if A ∈ Q10 , it can be canonically expanded as
∞
(1.27) A= bn 10n .
n=−f
In order to find the digits of A2 and A5 , we write

∞ ∞ ∞ ∞

A2 = (bn 5n )2n = cn 2n , A5 = (bn 2n )5n = dn 5n ,
n=−f n=−f n=−f n=−f
where the coefficients cn and dn in the respective canonical expansions

are obtained by reduction to the canonical form (Theorem 1.30). Thus
we have A = A2 , A5 , and from the rules for the sum, difference, and
product of g-adic and p-adic limits, we see that if B = B2 , B5 is a
second 10-adic number with B2 ∈ Q2 and B5 ∈ Q5 , then
A ± B = A2 ± B2 , A5 ± B5 and AB = A2 B2 , A5 B5 .
Conversely, let A2 ∈ Q2 and A5 ∈ Q5 be arbitrary. We will show that

there exists an A ∈ Q10 such that A = A2 , A5 . For p = 2, 5, let
(p)
{an } be a sequence of rational numbers such that Ap = lim a(p)
n in
n→∞
| · |p .
(p)
There is no reason why the sequence {an } should converge in
| · |q for q = p, and it may not even be bounded relative to | · |q . In
order to overcome this difficulty, we consider the sequences
5n 2n
e(2)
n = n n
, e(5)
n = n .
2 +5 2 + 5n
It is easy to see that

1 if p = q,
lim en = δpq in | · |q , where δpq =
(p)
n→∞ 0 if p = q.
(p)
It follows that there is an infinite subsequence ern such that

(q) Ap if p = q,
(p) (p)
lim a ern =
n→∞ n 0 if p = q.
Hence
(10) (10)
n ern = A2 , 0, and lim an ern = 0, A5 .
lim a(2) (2) (5) (5)
n→∞ n→∞
(2) (2) (5) (5)
Finally, we see that the sequence an = an ern + an ern converges to
A2 , A5 = A.
Exercises 47–50
Exercise 47. Prove that if g is not a prime, then | · |g is a pseudo-

norm on Q, i.e., it satisfies (1) and (3) of Definition 1.5 and (1.26).
Exercise 48. Prove that Q10 is not a field by displaying zero divi-
sors.
Exercise 49.* Look at the following sequence of integers:

6, 76, 376, 9376, 109376 . . . .
Exercises 47–50 51
(1) Prove that it can be continued in a unique way to obtain a 10-adic

integer α = . . . 109376 such that α2 = α.
(2) Prove that the equation x2 = x has 4 solutions in Z10 , namely
0, 1, α and β.
(3) Find the first 6 digits of β.
(4) Prove that Z10 ≈ Z5 ⊕ Z2 (direct product of groups).
Exercise 50. Prove that there is no relation of total order ≤ on Qp

possessing the following properties:
(1) if x ≤ y, then z + x ≤ z + y for any z;
(2) if 0 ≤ x and 0 ≤ y, then 0 ≤ xy.

STML 37 Prev

Uploaded by

Copyright:

Available Formats

STML 37 Prev

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

STML 37 Prev

Uploaded by

Copyright:

Available Formats

Chapter 1

Arithmetic of the p-adic

1.1. From Q to R; the concept of completion

Theorem 1.1 (Completion Theorem). Every metric space M can be

A = ({an }) and B = ({bn }) is deﬁned by the formula

(1.1) D(A, B) = lim d(an , bn ).

(1.2) d(r1 , r2 ) = |r1 − r2 |,

where m is a certain integer and the coeﬃcients or digits ak take the

.1, .1011, .10110111, .1011011101111, . . . ;

hence Q is not complete with respect to the Euclidean distance. On

Deﬁnition 1.2. Let (M, d) be a metric space, and R+ denotes the

B(a, r) = {x ∈ M | d(a, x) < r}.

The closed ball of radius r ∈ R+ centered at a ∈ M is the set

B(a, r) = {x ∈ M | d(a, x) ≤ r}.

A set A ⊂ M is called open if for any x ∈ M there exists an open

Practically, completion of a metric space is often obtained by a

Proposition 1.3. Let M be a complete metric space and let X be a

Example 1.4. The completion of an open interval (a, b) with respect

For other examples, see Exercise 5.

Exercise 1. Prove that any Cauchy sequence of rational numbers

Exercise 2. Prove that a number is rational if and only if its repre-

Exercise 3. Use the representation of real numbers as inﬁnite deci-

Exercise 4. Prove Proposition 1.3.

Exercise 6. Prove that a metric space is complete if and only if the

Exercise 7. Let {an } and {bn } be two Cauchy sequences in a metric

1.2. Normed ﬁelds

The norm is called trivial if 0 = 0 and x = 1 for all x = 0.

We shall denote this element by the same symbol n as the correspond-

Proposition 1.6. For any x, y ∈ F we have

Let d(x, y) = x − y . It follows immediately from Deﬁnition 1.5

Deﬁnition 1.7. A sequence {an } in F is said to be

• bounded if there is a constant C > 0 such that

It follows from the deﬁnition that any null sequence converges

The following is a simple but very useful result.

Proposition 1.8. x < 1 if and only if limn→∞ xn = 0.

Proof. Let x < 1. Since xn = x n , we obtain

i.e., limn→∞ xn = 0. Conversely, if x ≥ 1, then for all positive n

Deﬁnition 1.9. We say that two metrics d1 and d2 on F are equiv-

Proposition 1.10. Let · 1 and · 2 be two norms on a ﬁeld F .

Proof. Suppose · 1 ∼ · 2 . If · 1 is trivial, then by Exercise 12

Then by Exercise 13, xm

with · 2 and · 1 interchanged, so we also ﬁnd that

Now we will describe all norms on Q equivalent to the absolute

Proof. Suppose α ≤ 1. The ﬁrst two properties of the norm are

It will follow from Ostrowski’s Theorem (Theorem 1.50) that

The metric induced by a non-Archimedean norm is said to be an

Proof. (a) ⇒ (b). We will prove this implication by induction.

Observe that n = (n − 1) + 1 ≤ max{ n − 1 , 1} = 1.

(b) ⇒ (a). We have

This proposition helps explain the diﬀerence between Archimedean

We leave it to the reader to check that a norm is Archimedean if and

Proof. By the strong triangle inequality,

We leave the proof of the next rather surprising proposition to

We shall conclude this section by showing that an Archimedean

Proof. It follows from Exercise 13 that if · 1 ∼ · 2 , then for

Exercise 9. Prove that a ﬁeld does not contain zero divisors.

Exercise 11. Prove that in a normed ﬁeld the following assertions

Exercise 12. Prove that if · 1 ∼ · 2 and if · 1 is trivial, so is

Exercise 13. Prove that if · 1 ∼ · 2 , then x 1 < 1 if and only

The norm is called trivial if 0 = 0 and x = 1 for all x = 0.

Let d(x, y) = x − y. It follows immediately from Deﬁnition 1.5

Proof. Let x < 1. Since xn = xn , we obtain

Observe that n = (n − 1) + 1 ≤ max{n − 1, 1} = 1.

Exercise 13. Prove that if · 1 ∼ · 2 , then x1 < 1 if and only

where the a ∈ F belong to diﬀerent equivalent classes in F for diﬀerent

Theorem 1.19. F is a ﬁeld.

Proof. It is easy to check that F, with operations deﬁned as follows:

Now we extend the norm · from F to F.

Deﬁnition 1.20. For any A ∈ F put

Proposition 1.21. · is a norm on F.

(3) Similarly, we obtain A + B ≤ A + B.

Now we can deﬁne bounded, Cauchy, and null sequences in F

Theorem 1.22. F is complete with respect to the norm · , and F

This proves that F is dense in F. Now suppose {An } = {A1 , A2 , . . . }

is a null sequence in F. This implies that

ai ≡ ai0 ≡ ai0 ≡ ai (mod pi0 ),

i.e., ai ≡ ai (mod pi0 ). But this means exactly that

where d−m = 0 and bi ∈ {0, 1, 2, . . . , p − 1}, and represent the given

has a multiplicative inverse in Zp if and only if a0 = 0.