Pure Mathematics Research-Number Theory(5)
Kronecker-Weber theorems
We now state the
global
Kronecker-Weber theorem。
-
-
Let
be a finite abelian extension。 For each ramified prime
of
pick a prime
and let
be the completion of
at
the fact that
is Galois means that every
is ramified with the same ramification index; it makes no difference which
we pick
We have
so
is an abelian extension of
and the local Kronecker-Weber theorem implies that
for some
Let
put
this is a finite product
and let
We will show
which implies
The field
is a compositum of Galois extensions of
and is therefore Galois over
with
isomorphic to a subgroup of
hence abelian
as recalled below, the Galois group of a compositum
of Galois extensions
is isomorphic to a subgroup of the direct product of the
Let
be a prime of
lying above a ramified prime
as above, the completion
of
at
is a finite abelian extension of
since
is finite abelian, and we have
Let
be the maximal unramified extension of
in
Then
is totally ramified
and
is isomorphic to the inertia group
the
all coincide because
is abelian
It follows from that
since
and
is unramified, and that
since
is unramified。 Moreover, we have
since
is totally ramified, and it follows that
Now let
be the group generated by the union of the groups
for
Since
is abelian, we have
thus
Each inertia field
is unramified at
as is
So
is unramified, and therefore
Thus
and
as claimed and
To prove the local Kronecker-Weber theorem we first reduce to the case of cyclic extensions of prime-power degree。 Recall that if
and
are two Galois extensions of a field
then their compositum
is Galois over
with Galois group
The inclusion on the
is an equality if and only if
Conversely, if
then by defining
and
we have
with
and
and
It follows from the structure theorem for finite abelian groups that we may decompose any finite abelian extension
into a compositum
of linearly disjoint cyclic extensions
of prime-power degree。 If each
lies in a cyclotomic field
then so does
Indeed,
where
To prove the local Kronecker-Weber theorem it thus suffices to consider cyclic extensions
of prime power degree
There two distinct cases:
and
Kronecker-Weber theorem for
Let
be the maximal unramified extension of
in
then
for some
The extension
is totally ramified, and it must be tamely ramified, since the ramification index is a power of
we have
for some uniformizer
with
We may assume that
for some
since
is unramified: if
is the maximal ideal of
then the valuation
extends
with index
so
The field
lies in the compositum of
and
and we will show that both fields lie in a cyclotomic extension of
The extension
is unramified, since
for
so
is unramified and
for some
The field
is a compositum of abelian extensions, so
is abelian, and it contains the subexten-sion
which must be Galois
since it lies in an abelian extension
and totally ramified
since it is an Eisenstein extension
The field
con-tains
take ratios of roots of
and is totally ramified, but
is unramified
since
so we must have
Thus
It follows that
We then have
and may take
Let
Then
is a root of the Eisenstein polynomial
so the extension
is totally ramified of degree
and
is a uniformizer Let
The minimal polynomial of
is
which is Eisenstein, so
is also totally ramified of degree
and
is a uniformizer。 We have
so
is a unit in the ring of integers of
If we now put
then
and
so we can lift
to a root
of
in
We then have
so
and therefore
is a root of the minimal polynomial of
Since
is Galois, this implies that
and since
and
both have degree
the two fields coincide。
To complete the proof of the local Kronecker-Weber theorem, we need to address the case
Kummer theory
Let
be a positive integer and let
be a field of characteristic prime to
that contains a primitive
th root of unity
While we are specifically interested in the case where
is a local or global field, in this section
can be any field that satisfies these conditions。
For any
the field
is the splitting field of
over
the notation
denotes a particular
th root of
but it does not matter which root we pick because all the
th roots of
lie in
if
then
for some
and
The polynomial
is separable, since
is prime to the characteristic of
so L is a Galois extension of
and
is cyclic, since we have an injective homomorphism
This homomorphism is an isomorphism if and only if
is irreducible。
Kummer’s key observation is that the converse holds。 In order to prove this we firstrecall a basic (but often omitted) lemma from Galois theory。
-
Suppose not。 Let
with
and
minimal; we must have
the
nonzero, and the
distinct。 Choose
so
possible since
We have
for all
and the same applies to
which yields a shorter relation
where
with
which is nontrivial because
a contradiction。
The automorphism
is a linear transformation of
with characteristic polynomial
by
this must be its minimal polynomial, since
is linearly independent。 Therefore
is eigenvalue of
and the lemma follows。
is a special case of Hilbert’s Theorem 90, which re-places
with any element
of norm
for example。
Let
be a cyclic extension of degree
with
By
there exists an element
for which
We have
thus
is invariant under the action of
and therefore lies in
Moreover, the orbit
of
under the action of
has order
so
as desired。
Let
be a field with algebraic closure
let
be prime to the characteristic of
and assume
The
is the map
where
is any
th root of
in
If
and
are two nth roots of
then
so
is fixed by
and
so the value of
does not depend on the choice of
If
then
for all
so the Kummer pairing depends only on the image of
in
thus we may also view it as a pairing on
For each
if we pick an
th root
of
then the extension
will be non-trivial and some
must act nontrivially on
For this
we have
so
thus
is injective。
Now let
be a homomorphism, and put
and
Then
so
is a cyclic extension of degree
and
implies that
for some
If we put
and consider the homomorphisms
for
these homomorphisms are all distinct
because the
are distinct modulo
and
is injective
and they all have the same kernel and image as
their kernels have the same fixed field
because
contains all the
th roots of
There are
distinct isomorphisms
one of which corresponds to
and each corresponds to one of the
It follows that
for some
thus
is surjective。
Given a finite subgroup
of
we can choose
so that the images
of the
in
form a basis for the abelian group
this means
where
is the order of
in
For each
the fixed field of the kernel of
is a cyclic extension of
isomorphic to
as in the proof of
The fields
are linearly disjoint over
because the
correspond to independent generators of
and their compositum
has Galois group
an abelian group whose exponent divides
such fields
are called
-
of
Conversely, given an
-Kummer extension
we can iteratively apply
to put
in the form
with each
and
and the images of the
in
then generate a subgroup
corresponding to
as above。We thus have a 1-to-1 correspondence between finite subgroups of
and (finite)
-Kummer extensions of
this correspondence also extends to infinite subgroups provided we put a suitable topology on the groups
So far we have been assuming that
contains all the
th roots of unity。 To help handle situations where this is not necessarily the case, we rely on the following lemma。
Let
let
and let
be the subgroup of
generated by
The Kummer pairing induces a bilinear pairing
that is compatible with the Galois action of
In particular, we have
for all
and
the Galois action on
is by conjugation
lift
to
and conjugate there
but it is trivial because
is abelian
so
The isomorphism
induced by the Kummer pairing is injective,so
The local Kronecker-Weber theorem for
We are now ready to prove the local Kronecker-Weber theorem in the case
There are two obvious candidates for
namely, the cyclotomic field
which is an unramified extension of degree
and the index
subfield of the cyclotomic field
which is a totally ramified extension of degree
the
-cyclotomic polynomial
has degree
and remains irreducible over
If
is contained in the compositum of these two fields then
where
and the theorem holds。 Otherwise, the field
is a Galois extension of
with
for some
the first factor comes from the Galois group of
the second two factors come from the Galois group of
note
and the last factor comes from the fact that we are assuming
so
is nontrivial and must have order
for some
It follows that the abelian group
has a quotient isomorphic to
and the subfield of
corresponding to this quotient is an abelian extension of
with Galois group isomorphic
By
below, no such field exists。
To prove that
admits no
-extensions our strategy is to use Kummer theory to show that the corresponding subgroup of
given by
must have
-rank
and therefore cannot exist。 For an alternative proof that uses higher ramification groups instead of Kummer theory。
Suppose for the sake of contradiction that
is an extension of
with Galois group
Then
is linearly disjoint from
since the order of
is not divisible by
and
is a
-Kummer extension。 There is thus a subgroup
isomorphic to
for which
where
here we identify elements of
by representatives in
that are determined only up to
th powers
For any
the extension
is abelian, so by
we have
for all
where
is the isomorphism defined by
The field
is a totally tamely ramified extension of
of degree
with residue field
as shown in the proof of
we may take
as a uniformizer。
For each
we have
thus
for all
hence for al
for
this implies
Now
is determined only up to
th-powers, so after multiplying by
we may assume
and after multiplying by a suitable power of
we may assume
since the image of
generates the multiplicative group
of the residue field。
We may thus assume that
where
Each
can be written as a power series in
with integer coefficients in
and constant coefficient
We have
since
and
for integers
For
we can choose
so that for some integer
and
we have
For
we have
since each term in the sum is congruent to
modulo
here we are representing
as an integer in
Thus
and
We also have
As we showed for
above, any
can be written as
with
Each interior term in the binomial expansion of
other than leading
is a multiple of
with
and it follows that
Thus every element of
is congruent to
modulo
and as you will show on the problem set, the converse holds, that is,
We know from
that
so
and therefore
For
this is possible only if
for every
equivalently, for every
but then
and we must have
since
We have shown that every
is represented by an element
with
and therefore lies in the subgroup of
generated by
and
which is an abelian group of exponent
generated by
elements, hence isomorphic to a subgroup of
But this contradicts
In the proof of
above, the elements of
that lie in
are quite special。 For most
the extension
will not be abelian, even though the extensions
and
both are, and we typically will not have
consider
The key point is that we started with an abelian extension
so
is an abelian extension containing
this ensures that for
the fields
are abelian。
There is an alternative proof to
that is much more explicit。One can show that for
the field
admits exactly
cyclic extensions of degree
the unramified extension
and the extensions
for integers
This implies that
cannot have a
extension, since this would imply the existence of
cyclic extensions of degree
one for each index
subgroup of
For
there is an extension of
with Galois group isomorphic to
the cyclotomic field
so the proof we used for
will not work。 However we can apply a completely analogous argument。
The unramified cyclotomic field
has Galois group
and the totally ramified cyclotomic field
has Galois group
up to isomorphism
Let
If
is not contained in
then
and thus admits a quotient isomorphic to
or
By
below, no extension of
has either of these Galois groups, thus
must lie in
There are exactly
quadratic extensions of
it follows that no extension of
has Galois group
since this group has
subgroups of index
whose fixed fields would yield
distinct quadratic extension of
There are only finitely many extensions of
of any fixed degree
and these can be enumerated by considering Eisenstein polynomials in
of degrees dividing
up to an equivalence relation implied by Krasner’s lemma。 One finds that there are
quartic extensions of
of which
are cyclic。 It follows that no extension of
has Galois group
since this group has
subgroups whose fixed fields would yield
distinct cyclic quartic extensions of