Namespaces
Variants
Actions

Difference between revisions of "Class field theory"

From Encyclopedia of Mathematics
Jump to: navigation, search
(Importing text file)
 
m (fixed diagrams)
 
(3 intermediate revisions by 3 users not shown)
Line 1: Line 1:
The theory that gives a description of all Abelian extensions (finite Galois extensions having Abelian Galois groups) of a field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c0223701.png" /> that belongs to one of the following types: 1) <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c0223702.png" /> is an algebraic number field, i.e. a finite extension of the field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c0223703.png" />; 2) <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c0223704.png" /> is a finite extension of the field of rational <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c0223705.png" />-adic numbers <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c0223706.png" />; 3) <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c0223707.png" /> is a field of algebraic functions in one variable over a finite field; and 4) <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c0223708.png" /> is the field of formal power series over a finite field.
+
<!--
 +
c0223701.png
 +
$#A+1 = 215 n = 1
 +
$#C+1 = 215 : ~/encyclopedia/old_files/data/C022/C.0202370 Class field theory
 +
Automatically converted into TeX, above some diagnostics.
 +
Please remove this comment and the {{TEX|auto}} line below,
 +
if TeX found to be correct.
 +
-->
 +
 
 +
{{TEX|auto}}
 +
{{TEX|done}}
 +
 
 +
The theory that gives a description of all Abelian extensions (finite Galois extensions having Abelian Galois groups) of a field $  K $
 +
that belongs to one of the following types: 1) $  K $
 +
is an algebraic number field, i.e. a finite extension of the field $  \mathbf Q $;  
 +
2) $  K $
 +
is a finite extension of the field of rational $  p $-adic numbers $  \mathbf Q _ {p} $;  
 +
3) $  K $
 +
is a field of algebraic functions in one variable over a finite field; and 4) $  K $
 +
is the field of formal power series over a finite field.
  
 
The basic theorems in class field theory were formulated and proved in particular cases by L. Kronecker, H. Weber, D. Hilbert, and others (see also [[Algebraic number theory|Algebraic number theory]]).
 
The basic theorems in class field theory were formulated and proved in particular cases by L. Kronecker, H. Weber, D. Hilbert, and others (see also [[Algebraic number theory|Algebraic number theory]]).
Line 5: Line 24:
 
Fields of the types 2) and 4) are called local, while those of types 1) and 3) are called global. Correspondingly, one can speak of local and global class field theory.
 
Fields of the types 2) and 4) are called local, while those of types 1) and 3) are called global. Correspondingly, one can speak of local and global class field theory.
  
In local class field theory, each finite Abelian extension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c0223709.png" /> with Galois group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237010.png" /> is put into correspondence with the norm subgroup <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237011.png" /> of the multiplicative group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237012.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237013.png" />. The group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237014.png" /> completely determines the field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237015.png" />, and there exists a canonical isomorphism <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237016.png" /> (the main isomorphism of class field theory). The theory of formal groups (see [[#References|[1]]]) gives an explicit form of this isomorphism. Conversely, any open subgroup of finite index in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237017.png" /> is realized as a norm subgroup for a certain Abelian extension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237018.png" /> (the existence theorem).
+
In local class field theory, each finite Abelian extension $  L/K $
 +
with Galois group $  G( L/K  ) $
 +
is put into correspondence with the norm subgroup $  N _ {L/K} ( L  ^ {*} ) $
 +
of the multiplicative group $  K  ^ {*} $
 +
of $  K $.  
 +
The group $  N _ {L/K} ( L  ^ {*} ) $
 +
completely determines the field $  L $,  
 +
and there exists a canonical isomorphism $  \phi : G( L/K  ) \simeq K  ^ {*} /N _ {L/K} ( L  ^ {*} ) $ (the main isomorphism of class field theory). The theory of formal groups (see [[#References|[1]]]) gives an explicit form of this isomorphism. Conversely, any open subgroup of finite index in $  K  ^ {*} $
 +
is realized as a norm subgroup for a certain Abelian extension $  L $ (the existence theorem).
 +
 
 +
If  $  L $
 +
and  $  L _ {1} $
 +
are finite Abelian extensions of a field  $  K $,
 +
$  M = L \cap L _ {1} $
 +
and  $  N = L \cdot L _ {1} $,
 +
then
 +
 
 +
$$ \tag{1 }
 +
\left .
 +
\begin{array}{c}
 +
N _ {M/K} ( M  ^ {*} )  = \
 +
N _ {L/K} ( L  ^ {*} ) N _ {L _ {1}  /K } ( L _ {1}  ^ {*} ) ,
 +
\\
  
If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237019.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237020.png" /> are finite Abelian extensions of a field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237021.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237022.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237023.png" />, then
+
N _ {N/K} ( N  ^ {*} )  = \
 +
N _ {L/K} ( L  ^ {*} ) \cap N _ {L _ {1}  /K } ( L _ {1}  ^ {*} ).  
 +
\end{array}
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237024.png" /></td> <td valign="top" style="width:5%;text-align:right;">(1)</td></tr></table>
+
\right \}
 +
$$
  
The inclusion <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237025.png" /> holds if and only if
+
The inclusion $  L _ {1} \supseteq L $
 +
holds if and only if
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237026.png" /></td> </tr></table>
+
$$
 +
N _ {L/K} ( L  ^ {*} )  \supset  N _ {L _ {1}  /K } ( L _ {1}  ^ {*} ),
 +
$$
  
 
and in that case the diagram
 
and in that case the diagram
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237027.png" /></td> <td valign="top" style="width:5%;text-align:right;">(2)</td></tr></table>
+
$$ \tag{2 }
 +
 
 +
\begin{array}{ccc}
 +
G( L _ {1} /K  )  & \overset{\phi}{\simeq}  & K  ^ {*} /N _ {L _ {1}  /K } ( L  ^ {*} )  \\
 +
{\alpha } \downarrow  &{}  &\downarrow {\beta }  \\
 +
G( L/K  )  & \underset{\phi}{\simeq} &K  ^ {*} /N _ {L/K} ( L  ^ {*} ) \\
 +
\end{array}
 +
 
 +
$$
  
is commutative, where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237028.png" /> is obtained by restricting the automorphism from <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237029.png" /> to <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237030.png" />, while <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237031.png" /> is induced by the identity mapping <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237032.png" />. In particular, if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237033.png" /> is the maximal Abelian extension of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237034.png" />, then the Galois group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237035.png" /> is canonically isomorphic to the profinite completion of the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237036.png" />.
+
is commutative, where $  \alpha $
 +
is obtained by restricting the automorphism from $  L _ {1} $
 +
to $  L $,  
 +
while $  \beta $
 +
is induced by the identity mapping $  K  ^ {*} \rightarrow K  ^ {*} $.  
 +
In particular, if $  K  ^ {ab} $
 +
is the maximal Abelian extension of $  K $,  
 +
then the Galois group $  G( K  ^ {ab} /K  ) $
 +
is canonically isomorphic to the profinite completion of the group $  K  ^ {*} $.
  
The isomorphism <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237037.png" /> also gives a description of the sequence of ramification subgroups in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237038.png" />. For example, the extension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237039.png" /> is unramified if and only if the group of units <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237040.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237041.png" /> is contained in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237042.png" />. In that case the isomorphism <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237043.png" /> is completely determined by the fact that the Frobenius automorphism that generates the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237044.png" /> corresponds to the class <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237045.png" />, where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237046.png" /> is a prime element of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237047.png" />.
+
The isomorphism $  \phi $
 +
also gives a description of the sequence of ramification subgroups in $  G( L/K  ) $.  
 +
For example, the extension $  L/K $
 +
is unramified if and only if the group of units $  U( K  ) $
 +
of $  K $
 +
is contained in $  N _ {L/K} ( L  ^ {*} ) $.  
 +
In that case the isomorphism $  \phi $
 +
is completely determined by the fact that the Frobenius automorphism that generates the group $  G( L/K  ) $
 +
corresponds to the class $  \pi \cdot N _ {L/K} ( L  ^ {*} ) $,  
 +
where $  \pi $
 +
is a prime element of $  K $.
  
In the language of group cohomology the isomorphism <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237048.png" /> is interpreted as an isomorphism between the Tate cohomology groups:
+
In the language of group cohomology the isomorphism $  \phi $
 +
is interpreted as an isomorphism between the Tate cohomology groups:
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237049.png" /></td> </tr></table>
+
$$
 +
H  ^ {- 2} ( G( L/K  ), \mathbf Z )  \simeq  G( L/K  )
 +
$$
  
 
and
 
and
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237050.png" /></td> </tr></table>
+
$$
 +
H  ^ {0} ( G( L/K  ), L  ^ {*} )  = K  ^ {*} /N _ {L/K} ( L  ^ {*} ).
 +
$$
  
Moreover, let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237051.png" /> be an arbitrary finite Galois extension of local fields. Then for any integer <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237052.png" /> there is a canonical isomorphism <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237053.png" />:
+
Moreover, let $  L/K $
 +
be an arbitrary finite Galois extension of local fields. Then for any integer $  n $
 +
there is a canonical isomorphism $  \phi _ {n} $:
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237054.png" /></td> </tr></table>
+
$$
 +
H  ^ {n- 2} ( G( L/K  ), \mathbf Z )  \simeq  H  ^ {n} ( G( L/K  ), L  ^ {*} ).
 +
$$
  
If a tower of Galois fields <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237055.png" /> is given, then the inflation
+
If a tower of Galois fields $  M \supset L \supset K $
 +
is given, then the inflation
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237056.png" /></td> </tr></table>
+
$$
 +
\inf : H  ^ {2} ( G( L/K  ), L  ^ {*} )  \rightarrow  H  ^ {2} ( G( M/K  ), M  ^ {*} )
 +
$$
  
 
preserves the invariant (see [[Brauer group|Brauer group]]) and the restriction
 
preserves the invariant (see [[Brauer group|Brauer group]]) and the restriction
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237057.png" /></td> </tr></table>
+
$$
 +
\mathop{\rm res} : H  ^ {2} ( G( M/K  ), M  ^ {*} )  \rightarrow  H  ^ {2} ( G( M/L), M  ^ {*} )
 +
$$
 +
 
 +
multiplies the invariant by  $  [ L : K] $.
 +
If  $  \overline{K}\; $
 +
is the separable closure of  $  K $,
 +
the invariant defines a canonical isomorphism between the Brauer group of  $  K $,
 +
 
 +
$$
 +
\mathop{\rm Br} ( K  )  \simeq  H  ^ {2} ( G ( \overline{K}\; / K  ), \overline{ {K  ^ {*} }}\; ) ,
 +
$$
 +
 
 +
and  $  \mathbf Q / \mathbf Z $.
 +
 
 +
In global class field theory, the role of the multiplicative group is played by the idèle class group (cf. [[Idèle|Idèle]]). Let  $  L/K $
 +
be a finite Galois extension of global fields and let  $  I _ {L} $
 +
be the idèle group of the field  $  L $.
 +
The group  $  L  ^ {*} $
 +
is imbedded in  $  I _ {L} $
 +
as a discrete subgroup (it is called the group of principal idèles), while the quotient group  $  C _ {L} = I _ {L} / L  ^ {*} $,
 +
provided with the quotient topology, is called the idèle class group. It can be shown that  $  H  ^ {1} ( G( L/K  ), C _ {L} )= 1 $
 +
and  $  H  ^ {2} ( G( L/K  ), C _ {L} ) \simeq \mathbf Z /n \mathbf Z $,
 +
where  $  n = [ L :  K] $.  
 +
One has the canonical imbedding  $  \mathop{\rm inv} :  H  ^ {2} ( G( L/K  ), C _ {L} ) \rightarrow \mathbf Q / \mathbf Z $.
 +
As in local class field theory, for any integer  $  n $
 +
there is an isomorphism (the main isomorphism of global class field theory):
  
multiplies the invariant by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237058.png" />. If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237059.png" /> is the separable closure of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237060.png" />, the invariant defines a canonical isomorphism between the Brauer group of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237061.png" />,
+
$$
 +
\psi _ {n} : H  ^ {n- 2} ( G( L/K  ), \mathbf Z )  \simeq  H  ^ {n} ( G( L/K  ), C _ {L} ).
 +
$$
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237062.png" /></td> </tr></table>
+
For an Abelian extension  $  L/K $,
 +
the isomorphism  $  \psi _ {0} $
 +
reduces to the isomorphism  $  \psi : G( L/K  ) \simeq C _ {K} /N _ {L/K} ( C _ {L} ) $.
 +
The norm subgroup  $  N _ {L/K} ( C _ {L} ) $
 +
uniquely determines the field  $  L $,
 +
and, conversely, any open subgroup of finite index in  $  C _ {K} $
 +
is a norm subgroup for some finite Abelian extension  $  L $ (the global existence theorem). Relationships analogous to (1) and (2) are also valid for global fields. If  $  K  ^ {ab} $
 +
is the maximal Abelian extension of a field  $  K $,
 +
then in the function field case the group  $  G( K  ^ {ab} /K  ) $
 +
is isomorphic to the profinite completion of the group  $  C _ {K} $,
 +
while in the number field case the group  $  G( K  ^ {ab} /K  ) $
 +
is isomorphic to the quotient group of the group  $  C _ {K} $
 +
by the connected component.
  
and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237063.png" />.
+
The isomorphisms  $  \phi _ {n} $
 +
and $  \psi _ {n} $
 +
are compatible. If  $  L/K $
 +
is a finite Galois extension of global fields,  $  L _ {v} $
 +
is the completion of  $  L $
 +
with respect to some valuation  $  v $
 +
and  $  K _ {v} $
 +
is the completion of  $  K $
 +
with respect to the restriction of  $  v $
 +
on  $  K $,
 +
then there exists a commutative diagram
  
In global class field theory, the role of the multiplicative group is played by the idèle class group (cf. [[Idèle|Idèle]]). Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237064.png" /> be a finite Galois extension of global fields and let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237065.png" /> be the idèle group of the field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237066.png" />. The group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237067.png" /> is imbedded in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237068.png" /> as a discrete subgroup (it is called the group of principal idèles), while the quotient group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237069.png" />, provided with the quotient topology, is called the idèle class group. It can be shown that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237070.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237071.png" />, where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237072.png" />. One has the canonical imbedding <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237073.png" />. As in local class field theory, for any integer <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237074.png" /> there is an isomorphism (the main isomorphism of global class field theory):
+
$$ \tag{3 }
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237075.png" /></td> </tr></table>
+
\begin{array}{ccc}
 +
H  ^ {n- 2} ( G( L/K  ), \mathbf Z )  & \overset{\psi _ n}{\simeq}    &H  ^ {n} ( G( L/K  ), C _ {L} )  \\
 +
\mathop{\rm cores} \uparrow  &{}  &\uparrow {f }  \\
 +
H  ^ {n- 2} ( G( L _ {v} / K _ {v} ), \mathbf Z )  & \underset{\psi_{n}}{\simeq} &H  ^ {n} ( G( L _ {v} / K _ {v} ), L _ {v}  ^ {*} ) ,  \\
 +
\end{array}
  
For an Abelian extension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237076.png" />, the isomorphism <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237077.png" /> reduces to the isomorphism <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237078.png" />. The norm subgroup <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237079.png" /> uniquely determines the field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237080.png" />, and, conversely, any open subgroup of finite index in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237081.png" /> is a norm subgroup for some finite Abelian extension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237082.png" /> (the global existence theorem). Relationships analogous to (1) and (2) are also valid for global fields. If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237083.png" /> is the maximal Abelian extension of a field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237084.png" />, then in the function field case the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237085.png" /> is isomorphic to the profinite completion of the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237086.png" />, while in the number field case the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237087.png" /> is isomorphic to the quotient group of the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237088.png" /> by the connected component.
+
$$
  
The isomorphisms <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237089.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237090.png" /> are compatible. If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237091.png" /> is a finite Galois extension of global fields, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237092.png" /> is the completion of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237093.png" /> with respect to some valuation <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237094.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237095.png" /> is the completion of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237096.png" /> with respect to the restriction of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237097.png" /> on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237098.png" />, then there exists a commutative diagram
+
where the mapping  $  f $
 +
is induced by the imbedding  $  L _ {v}  ^ {*} \rightarrow I _ {L} \rightarrow C _ {L} $
 +
and the co-restriction mapping cores. For  $  n = 0 $,  
 +
(3) gives the commutative diagram
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c02237099.png" /></td> <td valign="top" style="width:5%;text-align:right;">(3)</td></tr></table>
+
$$ \tag{4 }
  
where the mapping <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370100.png" /> is induced by the imbedding <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370101.png" /> and the co-restriction mapping cores. For <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370102.png" />, (3) gives the commutative diagram
+
\begin{array}{ccc}
 +
G( L/K  )/[ G( L/K  ), G( L/K  )]  & \overset{\psi}{\simeq}  &C _ {K} /N _ {L/K} ( C _ {L} )  \\
 +
\uparrow  &{}  &\uparrow  \\
 +
G( L _ {v} / K _ {v} )/[ G( L _ {v} /K _ {v} ), G( L _ {v} / K _ {v} )]  & \underset{\phi}{\simeq} &K _ {v}  ^ {*} / N _ {L _ {v}  /K _ {v} } ( L _ {v}  ^ {*} ).  \\
 +
\end{array}
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370103.png" /></td> <td valign="top" style="width:5%;text-align:right;">(4)</td></tr></table>
+
$$
  
The diagram (4) enables one to obtain a decomposition law of prime divisors of the field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370104.png" /> in the Abelian extension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370105.png" />. That is, a prime divisor <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370106.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370107.png" /> is unramified (splits completely) in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370108.png" /> if and only if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370109.png" /> (correspondingly, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370110.png" />).
+
The diagram (4) enables one to obtain a decomposition law of prime divisors of the field $  K $
 +
in the Abelian extension $  L/K $.  
 +
That is, a prime divisor $  \mathfrak c $
 +
of $  K $
 +
is unramified (splits completely) in $  L $
 +
if and only if $  U( K _ {\mathfrak c }  ) \subset  N _ {L/K} ( C _ {L} ) $(
 +
correspondingly, $  K  ^ {*} \subset  N _ {L/K} ( C _ {L} ) $).
  
If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370111.png" /> is a prime divisor of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370112.png" /> that is unramified in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370113.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370114.png" /> is the valuation of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370115.png" /> corresponding to <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370116.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370117.png" /> is a prime element of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370118.png" />, then the Artin symbol
+
If $  \mathfrak c $
 +
is a prime divisor of $  K $
 +
that is unramified in $  L $,  
 +
$  v $
 +
is the valuation of $  K $
 +
corresponding to $  \mathfrak c $
 +
and $  \pi $
 +
is a prime element of $  K _ {v} $,  
 +
then the Artin symbol
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370119.png" /></td> </tr></table>
+
$$
 +
\left ( \frac{L/K}{\mathfrak c}
 +
\right )  = \psi  ^ {- 1} ( \pi )  \in  G( L/K  )
 +
$$
  
is defined and only depends on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370120.png" />. It is the Frobenius automorphism in the decomposition subgroup of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370121.png" />. According to Chebotarev's density theorem, any element of the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370122.png" /> has the form
+
is defined and only depends on $  \mathfrak c $.  
 +
It is the Frobenius automorphism in the decomposition subgroup of $  v $.  
 +
According to Chebotarev's density theorem, any element of the group $  G( L/K  ) $
 +
has the form
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370123.png" /></td> </tr></table>
+
$$
 +
\left ( \frac{L/K}{\mathfrak c}
 +
\right )
 +
$$
  
for an infinite number of prime divisors <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370124.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370125.png" />.
+
for an infinite number of prime divisors $  \mathfrak c $
 +
of $  K $.
  
For example, the maximal unramified Abelian extension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370126.png" /> of a number field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370127.png" /> (called the Hilbert class field) is a field whose norm subgroup coincides with the image under the projection <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370128.png" /> of the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370129.png" />, where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370130.png" /> runs through all points of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370131.png" />. The group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370132.png" /> is canonically isomorphic to the class group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370133.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370134.png" />, which gives the important isomorphism <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370135.png" />. In particular, there are no unramified Abelian extensions of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370136.png" /> if and only if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370137.png" /> has class number one.
+
For example, the maximal unramified Abelian extension $  F $
 +
of a number field $  K $ (called the Hilbert class field) is a field whose norm subgroup coincides with the image under the projection $  I _ {K} \rightarrow C _ {K} $
 +
of the group $  K  ^ {*} \prod _ {v} U( K _ {v} ) $,  
 +
where $  v $
 +
runs through all points of $  K $.  
 +
The group $  I _ {K} /K  ^ {*} \prod _ {v} U( K _ {v} ) $
 +
is canonically isomorphic to the class group $  \mathop{\rm Cl} _ {K} $
 +
of $  K $,  
 +
which gives the important isomorphism $  G( F/K  ) \simeq  \mathop{\rm Cl} _ {K} $.  
 +
In particular, there are no unramified Abelian extensions of $  K $
 +
if and only if $  K $
 +
has class number one.
  
The type of decomposition for a prime divisor <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370138.png" /> of the field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370139.png" /> in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370140.png" /> is completely determined by the class of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370141.png" /> in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370142.png" />. In particular, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370143.png" /> splits completely if and only if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370144.png" /> is principal. All divisors of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370145.png" /> become principal divisors in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370146.png" />.
+
The type of decomposition for a prime divisor $  \mathfrak c $
 +
of the field $  K $
 +
in $  F $
 +
is completely determined by the class of $  \mathfrak c $
 +
in $  \mathop{\rm Cl} _ {K} $.  
 +
In particular, $  \mathfrak c $
 +
splits completely if and only if $  \mathfrak c $
 +
is principal. All divisors of $  K $
 +
become principal divisors in $  F $.
  
 
Just as class field theory for unramified Abelian extensions can be explained in terms of the divisor class group and its subgroups, so can arbitrary Abelian extensions be characterized by means of ray class groups with respect to suitable modules (see [[Algebraic number theory|Algebraic number theory]]). There are also generalizations of class field theory to the case of infinite Galois extensions [[#References|[4]]].
 
Just as class field theory for unramified Abelian extensions can be explained in terms of the divisor class group and its subgroups, so can arbitrary Abelian extensions be characterized by means of ray class groups with respect to suitable modules (see [[Algebraic number theory|Algebraic number theory]]). There are also generalizations of class field theory to the case of infinite Galois extensions [[#References|[4]]].
Line 84: Line 273:
  
 
====References====
 
====References====
<table><TR><TD valign="top">[1]</TD> <TD valign="top">  J.W.S. Cassels (ed.)  A. Fröhlich (ed.) , ''Algebraic number theory'' , Acad. Press  (1967)  pp. Chapt. VI</TD></TR><TR><TD valign="top">[2]</TD> <TD valign="top">  A. Weil,  "Basic number theory" , Springer  (1973)</TD></TR><TR><TD valign="top">[3]</TD> <TD valign="top">  H. Koch,  "Galoissche Theorie der <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370147.png" />-Erweiterungen" , Deutsch. Verlag Wissenschaft.  (1970)</TD></TR><TR><TD valign="top">[4]</TD> <TD valign="top">  L.V. Kuz'min,  "Homotopy of profinite groups, the Schur multiplicator and class field theory"  ''Izv. Akad. Nauk SSSR Ser. Mat.'' , '''33''' :  6  (1969)  pp. 1220–1254  (In Russian)</TD></TR></table>
+
<table><TR><TD valign="top">[1]</TD> <TD valign="top">  J.W.S. Cassels (ed.)  A. Fröhlich (ed.) , ''Algebraic number theory'' , Acad. Press  (1967)  pp. Chapt. VI</TD></TR><TR><TD valign="top">[2]</TD> <TD valign="top">  A. Weil,  "Basic number theory" , Springer  (1973)</TD></TR><TR><TD valign="top">[3]</TD> <TD valign="top">  H. Koch,  "Galoissche Theorie der $p$-Erweiterungen" , Deutsch. Verlag Wissenschaft.  (1970)</TD></TR><TR><TD valign="top">[4]</TD> <TD valign="top">  L.V. Kuz'min,  "Homotopy of profinite groups, the Schur multiplicator and class field theory"  ''Izv. Akad. Nauk SSSR Ser. Mat.'' , '''33''' :  6  (1969)  pp. 1220–1254  (In Russian)</TD></TR></table>
 
 
 
 
  
 
====Comments====
 
====Comments====
Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370148.png" /> be the ring of integers of the global field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370149.png" />. Then the class group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370150.png" /> is the [[Divisor class group|divisor class group]] of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370151.png" />; i.e. it is the group of classes of ideals of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370152.png" /> modulo principal ideals.
+
Let $  A $
 +
be the ring of integers of the global field $  K $.  
 +
Then the class group $  \mathop{\rm Cl} _ {K} $
 +
is the [[Divisor class group|divisor class group]] of $  A $;  
 +
i.e. it is the group of classes of ideals of $  A $
 +
modulo principal ideals.
  
The fact that the divisors of a field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370153.png" /> become principal divisors in its maximal unramified Abelian extension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370154.png" /> is called the principal ideal theorem.
+
The fact that the divisors of a field $  K $
 +
become principal divisors in its maximal unramified Abelian extension $  F $
 +
is called the principal ideal theorem.
  
 
Two excellent modern up-to-date books on class field theory are [[#References|[a1]]] and [[#References|[a2]]]. The latter also discusses the relations between the idèle-theoretic formulation of class field theory and ray class groups.
 
Two excellent modern up-to-date books on class field theory are [[#References|[a1]]] and [[#References|[a2]]]. The latter also discusses the relations between the idèle-theoretic formulation of class field theory and ray class groups.
  
The Kronecker–Weber theorem states that every Abelian finite extension of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370155.png" /> is contained in some <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370156.png" /> where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370157.png" /> is a primitive <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370158.png" />-th root of unity (i.e. <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370159.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370160.png" /> for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370161.png" />). Kronecker also conjectured that every Abelian extension of an imaginary quadratic field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370162.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370163.png" />, is contained in an extension generated by the torsion points of an elliptic curve with complex multiplication. This was proved by T. Takagi [[#References|[a3]]]. Its analogue for local fields is the Lubin–Tate theorem, stating that the torsion points of a Lubin–Tate formal group over the ring of integers <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370164.png" /> of a local field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370165.png" /> together with the maximal unramified extension of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370166.png" /> generate the maximal Abelian extension of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370167.png" /> [[#References|[a4]]]. These formal groups can be used to give very explicit descriptions of the local reciprocity mappings <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370168.png" />, cf. also [[#References|[a5]]]. The Lubin–Tate formal groups are analogous to elliptic curves with complex multiplication in that they have maximally large endomorphism rings.
+
The Kronecker–Weber theorem states that every Abelian finite extension of $  \mathbf Q $
 +
is contained in some $  \mathbf Q ( \xi _ {n} ) $
 +
where $  \xi _ {n} $
 +
is a primitive $  n $-th root of unity (i.e. $  \xi _ {n}  ^ {n} = 1 $
 +
and $  \xi _ {n}  ^ {m} \neq 1 $
 +
for $  m < n $).  
 +
Kronecker also conjectured that every Abelian extension of an imaginary quadratic field $  \mathbf Q ( \sqrt{ - d} ) $,  
 +
$  d \in \mathbf N $,  
 +
is contained in an extension generated by the torsion points of an elliptic curve with complex multiplication. This was proved by T. Takagi [[#References|[a3]]]. Its analogue for local fields is the Lubin–Tate theorem, stating that the torsion points of a Lubin–Tate formal group over the ring of integers $  A $
 +
of a local field $  K $
 +
together with the maximal unramified extension of $  K $
 +
generate the maximal Abelian extension of $  K $[[#References|[a4]]]. These formal groups can be used to give very explicit descriptions of the local reciprocity mappings $  K  ^ {*} \rightarrow  \mathop{\rm Gal} ( K  ^ {ab} / K ) $,  
 +
cf. also [[#References|[a5]]]. The Lubin–Tate formal groups are analogous to elliptic curves with complex multiplication in that they have maximally large endomorphism rings.
  
The mapping <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370169.png" /> sets up a one-to-one correspondence between the finite Abelian extensions <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370170.png" /> and the closed subgroups of finite index in the idèle class group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370171.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370172.png" /> (cf. [[Idèle|Idèle]] for the topology on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370173.png" />). This is often called the existence theorem of class field theory. If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370174.png" /> is associated to a subgroup <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370175.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370176.png" />, then <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370177.png" /> is called the class field of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370178.png" />. Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370179.png" /> be a formal product of prime divisors of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370180.png" /> such that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370181.png" /> for all <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370182.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370183.png" /> for almost-all <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370184.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370185.png" /> or 1 for the infinite primes of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370186.png" />. Such a formal product is called a positive divisor or cycle. For each prime divisor <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370187.png" />, let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370188.png" /> be the local field obtained by completion with respect to the valuation defined by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370189.png" /> and let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370190.png" /> be its ring of integers. For each finite prime, let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370191.png" /> for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370192.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370193.png" />, the group of units of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370194.png" />. In addition, for the infinite primes <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370195.png" /> define <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370196.png" />, the positive reals, if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370197.png" /> is real, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370198.png" /> if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370199.png" /> is real, and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370200.png" /> if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370201.png" /> is complex. Given a positive divisor <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370202.png" />, a corresponding subgroup <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370203.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370204.png" /> is defined by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370205.png" /> where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370206.png" /> is the subgroup of the group of idèles, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370207.png" />, defined by
+
The mapping $  L \rightarrow N _ {L/K} C _ {L} \subset  C _ {K} $
 +
sets up a one-to-one correspondence between the finite Abelian extensions $  L / K $
 +
and the closed subgroups of finite index in the idèle class group $  C _ {K} $
 +
of $  K $ (cf. [[Idèle|Idèle]] for the topology on $  C _ {K} $).  
 +
This is often called the existence theorem of class field theory. If $  L / K $
 +
is associated to a subgroup $  N $
 +
of $  C _ {K} $,  
 +
then $  L / K $
 +
is called the class field of $  N $.  
 +
Let $  \mathfrak m = \prod _ {\mathfrak p} \mathfrak p ^ {n _ {\mathfrak p} } $
 +
be a formal product of prime divisors of $  K $
 +
such that $  n _ {\mathfrak p} \geq  0 $
 +
for all $  \mathfrak p $,  
 +
$  n _ {\mathfrak p} = 0 $
 +
for almost-all $  \mathfrak p $
 +
and $  n _ {\mathfrak p} = 0 $
 +
or 1 for the infinite primes of $  K $.  
 +
Such a formal product is called a positive divisor or cycle. For each prime divisor $  \mathfrak p $,  
 +
let $  K _ {\mathfrak p} $
 +
be the local field obtained by completion with respect to the valuation defined by $  \mathfrak p $
 +
and let $  A _ {\mathfrak p} $
 +
be its ring of integers. For each finite prime, let $  U _ {\mathfrak p}  ^ {n} = \{ {x \in A _ {\mathfrak p} } : {x \equiv 1  \mathop{\rm mod}  \mathfrak p  ^ {n} } \} $
 +
for  $  n > 0 $
 +
and $  U _ {\mathfrak p}  ^ {0} = U _ {\mathfrak p} = A _ {\mathfrak p}  ^ {*} $,  
 +
the group of units of $  A _ {\mathfrak p} $.  
 +
In addition, for the infinite primes $  \mathfrak p $
 +
define $  U _ {\mathfrak p}  ^ {1} = \mathbf R  ^ {+} $,  
 +
the positive reals, if $  \mathfrak p $
 +
is real, $  U _ {\mathfrak p}  ^ {0} = \mathbf R  ^ {*} $
 +
if $  \mathfrak p $
 +
is real, and $  U _ {\mathfrak p}  ^ {0} = U _ {\mathfrak p}  ^ {1} = \mathbf C  ^ {*} $
 +
if $  \mathfrak p $
 +
is complex. Given a positive divisor $  \mathfrak n $,  
 +
a corresponding subgroup $  C _ {K}  ^ {\mathfrak n} $
 +
of $  C _ {K} $
 +
is defined by $  C _ {k}  ^ {\mathfrak n} = I _ {K}  ^ {\mathfrak n} K  ^ {*} / K  ^ {*} $
 +
where $  I _ {K}  ^ {\mathfrak n} $
 +
is the subgroup of the group of idèles, $  I _ {K} $,  
 +
defined by
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370208.png" /></td> </tr></table>
+
$$
 +
I _ {K}  ^ {\mathfrak n} =
 +
\{ {\alpha \in I _ {K} } : {\alpha _ {\mathfrak p} \in U _ {\mathfrak p} ^ {n _ {\mathfrak p} }  \textrm{ for  all  }  \mathfrak p } \}
 +
$$
  
The subgroup <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370209.png" /> is called a congruence subgroup; more precisely, it is the congruence subgroup <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370210.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370211.png" />. The corresponding class field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370212.png" />, i.e. the Abelian extension such that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370213.png" />, is called the ray class field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370214.png" />. Of particular interest is the ray class field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370215.png" />, which is the Hilbert class field, since <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/c/c022/c022370/c022370216.png" /> is clearly isomorphic to the ideal class group of all ideals modulo principal ideals.
+
The subgroup $  C _ {K}  ^ {\mathfrak n} $
 +
is called a congruence subgroup; more precisely, it is the congruence subgroup $  \mathop{\rm mod}  \mathfrak n $
 +
of $  C _ {K} $.  
 +
The corresponding class field $  K  ^ {\mathfrak n} / K $,  
 +
i.e. the Abelian extension such that $  N _ {K  ^ {\mathfrak n}  / K } C _ {K  ^ {\mathfrak n}  } = C _ {K}  ^ {\mathfrak n} $,  
 +
is called the ray class field $  \mathop{\rm mod}  \mathfrak n $.  
 +
Of particular interest is the ray class field $  \mathop{\rm mod}  1 $,  
 +
which is the Hilbert class field, since $  C _ {K} / C _ {K}  ^ {1} $
 +
is clearly isomorphic to the ideal class group of all ideals modulo principal ideals.
  
 
====References====
 
====References====
 
<table><TR><TD valign="top">[a1]</TD> <TD valign="top">  K. Iwasawa,  "Local class field theory" , Oxford Univ. Press  (1986)</TD></TR><TR><TD valign="top">[a2]</TD> <TD valign="top">  J. Neukirch,  "Class field theory" , Springer  (1986)  pp. Chapt. 4, Sect. 8</TD></TR><TR><TD valign="top">[a3]</TD> <TD valign="top">  T. Takagi,  "Ueber eine Theorie des relativ-abelschen Zahlkörpers"  ''J. Coll. Sci. Imp. Univ. Tokyo'' , '''41'''  (1920)  pp. 1–132</TD></TR><TR><TD valign="top">[a4]</TD> <TD valign="top">  J. Lubin,  J. Tate,  "Formal complex multiplication in local fields"  ''Ann. of Math.'' , '''81'''  (1965)  pp. 380–387</TD></TR><TR><TD valign="top">[a5]</TD> <TD valign="top">  M. Hazewinkel,  "Local class field theory is easy"  ''Adv. in Math.'' , '''18'''  (1975)  pp. 148–181</TD></TR></table>
 
<table><TR><TD valign="top">[a1]</TD> <TD valign="top">  K. Iwasawa,  "Local class field theory" , Oxford Univ. Press  (1986)</TD></TR><TR><TD valign="top">[a2]</TD> <TD valign="top">  J. Neukirch,  "Class field theory" , Springer  (1986)  pp. Chapt. 4, Sect. 8</TD></TR><TR><TD valign="top">[a3]</TD> <TD valign="top">  T. Takagi,  "Ueber eine Theorie des relativ-abelschen Zahlkörpers"  ''J. Coll. Sci. Imp. Univ. Tokyo'' , '''41'''  (1920)  pp. 1–132</TD></TR><TR><TD valign="top">[a4]</TD> <TD valign="top">  J. Lubin,  J. Tate,  "Formal complex multiplication in local fields"  ''Ann. of Math.'' , '''81'''  (1965)  pp. 380–387</TD></TR><TR><TD valign="top">[a5]</TD> <TD valign="top">  M. Hazewinkel,  "Local class field theory is easy"  ''Adv. in Math.'' , '''18'''  (1975)  pp. 148–181</TD></TR></table>

Latest revision as of 07:48, 13 February 2024


The theory that gives a description of all Abelian extensions (finite Galois extensions having Abelian Galois groups) of a field $ K $ that belongs to one of the following types: 1) $ K $ is an algebraic number field, i.e. a finite extension of the field $ \mathbf Q $; 2) $ K $ is a finite extension of the field of rational $ p $-adic numbers $ \mathbf Q _ {p} $; 3) $ K $ is a field of algebraic functions in one variable over a finite field; and 4) $ K $ is the field of formal power series over a finite field.

The basic theorems in class field theory were formulated and proved in particular cases by L. Kronecker, H. Weber, D. Hilbert, and others (see also Algebraic number theory).

Fields of the types 2) and 4) are called local, while those of types 1) and 3) are called global. Correspondingly, one can speak of local and global class field theory.

In local class field theory, each finite Abelian extension $ L/K $ with Galois group $ G( L/K ) $ is put into correspondence with the norm subgroup $ N _ {L/K} ( L ^ {*} ) $ of the multiplicative group $ K ^ {*} $ of $ K $. The group $ N _ {L/K} ( L ^ {*} ) $ completely determines the field $ L $, and there exists a canonical isomorphism $ \phi : G( L/K ) \simeq K ^ {*} /N _ {L/K} ( L ^ {*} ) $ (the main isomorphism of class field theory). The theory of formal groups (see [1]) gives an explicit form of this isomorphism. Conversely, any open subgroup of finite index in $ K ^ {*} $ is realized as a norm subgroup for a certain Abelian extension $ L $ (the existence theorem).

If $ L $ and $ L _ {1} $ are finite Abelian extensions of a field $ K $, $ M = L \cap L _ {1} $ and $ N = L \cdot L _ {1} $, then

$$ \tag{1 } \left . \begin{array}{c} N _ {M/K} ( M ^ {*} ) = \ N _ {L/K} ( L ^ {*} ) N _ {L _ {1} /K } ( L _ {1} ^ {*} ) , \\ N _ {N/K} ( N ^ {*} ) = \ N _ {L/K} ( L ^ {*} ) \cap N _ {L _ {1} /K } ( L _ {1} ^ {*} ). \end{array} \right \} $$

The inclusion $ L _ {1} \supseteq L $ holds if and only if

$$ N _ {L/K} ( L ^ {*} ) \supset N _ {L _ {1} /K } ( L _ {1} ^ {*} ), $$

and in that case the diagram

$$ \tag{2 } \begin{array}{ccc} G( L _ {1} /K ) & \overset{\phi}{\simeq} & K ^ {*} /N _ {L _ {1} /K } ( L ^ {*} ) \\ {\alpha } \downarrow &{} &\downarrow {\beta } \\ G( L/K ) & \underset{\phi}{\simeq} &K ^ {*} /N _ {L/K} ( L ^ {*} ) \\ \end{array} $$

is commutative, where $ \alpha $ is obtained by restricting the automorphism from $ L _ {1} $ to $ L $, while $ \beta $ is induced by the identity mapping $ K ^ {*} \rightarrow K ^ {*} $. In particular, if $ K ^ {ab} $ is the maximal Abelian extension of $ K $, then the Galois group $ G( K ^ {ab} /K ) $ is canonically isomorphic to the profinite completion of the group $ K ^ {*} $.

The isomorphism $ \phi $ also gives a description of the sequence of ramification subgroups in $ G( L/K ) $. For example, the extension $ L/K $ is unramified if and only if the group of units $ U( K ) $ of $ K $ is contained in $ N _ {L/K} ( L ^ {*} ) $. In that case the isomorphism $ \phi $ is completely determined by the fact that the Frobenius automorphism that generates the group $ G( L/K ) $ corresponds to the class $ \pi \cdot N _ {L/K} ( L ^ {*} ) $, where $ \pi $ is a prime element of $ K $.

In the language of group cohomology the isomorphism $ \phi $ is interpreted as an isomorphism between the Tate cohomology groups:

$$ H ^ {- 2} ( G( L/K ), \mathbf Z ) \simeq G( L/K ) $$

and

$$ H ^ {0} ( G( L/K ), L ^ {*} ) = K ^ {*} /N _ {L/K} ( L ^ {*} ). $$

Moreover, let $ L/K $ be an arbitrary finite Galois extension of local fields. Then for any integer $ n $ there is a canonical isomorphism $ \phi _ {n} $:

$$ H ^ {n- 2} ( G( L/K ), \mathbf Z ) \simeq H ^ {n} ( G( L/K ), L ^ {*} ). $$

If a tower of Galois fields $ M \supset L \supset K $ is given, then the inflation

$$ \inf : H ^ {2} ( G( L/K ), L ^ {*} ) \rightarrow H ^ {2} ( G( M/K ), M ^ {*} ) $$

preserves the invariant (see Brauer group) and the restriction

$$ \mathop{\rm res} : H ^ {2} ( G( M/K ), M ^ {*} ) \rightarrow H ^ {2} ( G( M/L), M ^ {*} ) $$

multiplies the invariant by $ [ L : K] $. If $ \overline{K}\; $ is the separable closure of $ K $, the invariant defines a canonical isomorphism between the Brauer group of $ K $,

$$ \mathop{\rm Br} ( K ) \simeq H ^ {2} ( G ( \overline{K}\; / K ), \overline{ {K ^ {*} }}\; ) , $$

and $ \mathbf Q / \mathbf Z $.

In global class field theory, the role of the multiplicative group is played by the idèle class group (cf. Idèle). Let $ L/K $ be a finite Galois extension of global fields and let $ I _ {L} $ be the idèle group of the field $ L $. The group $ L ^ {*} $ is imbedded in $ I _ {L} $ as a discrete subgroup (it is called the group of principal idèles), while the quotient group $ C _ {L} = I _ {L} / L ^ {*} $, provided with the quotient topology, is called the idèle class group. It can be shown that $ H ^ {1} ( G( L/K ), C _ {L} )= 1 $ and $ H ^ {2} ( G( L/K ), C _ {L} ) \simeq \mathbf Z /n \mathbf Z $, where $ n = [ L : K] $. One has the canonical imbedding $ \mathop{\rm inv} : H ^ {2} ( G( L/K ), C _ {L} ) \rightarrow \mathbf Q / \mathbf Z $. As in local class field theory, for any integer $ n $ there is an isomorphism (the main isomorphism of global class field theory):

$$ \psi _ {n} : H ^ {n- 2} ( G( L/K ), \mathbf Z ) \simeq H ^ {n} ( G( L/K ), C _ {L} ). $$

For an Abelian extension $ L/K $, the isomorphism $ \psi _ {0} $ reduces to the isomorphism $ \psi : G( L/K ) \simeq C _ {K} /N _ {L/K} ( C _ {L} ) $. The norm subgroup $ N _ {L/K} ( C _ {L} ) $ uniquely determines the field $ L $, and, conversely, any open subgroup of finite index in $ C _ {K} $ is a norm subgroup for some finite Abelian extension $ L $ (the global existence theorem). Relationships analogous to (1) and (2) are also valid for global fields. If $ K ^ {ab} $ is the maximal Abelian extension of a field $ K $, then in the function field case the group $ G( K ^ {ab} /K ) $ is isomorphic to the profinite completion of the group $ C _ {K} $, while in the number field case the group $ G( K ^ {ab} /K ) $ is isomorphic to the quotient group of the group $ C _ {K} $ by the connected component.

The isomorphisms $ \phi _ {n} $ and $ \psi _ {n} $ are compatible. If $ L/K $ is a finite Galois extension of global fields, $ L _ {v} $ is the completion of $ L $ with respect to some valuation $ v $ and $ K _ {v} $ is the completion of $ K $ with respect to the restriction of $ v $ on $ K $, then there exists a commutative diagram

$$ \tag{3 } \begin{array}{ccc} H ^ {n- 2} ( G( L/K ), \mathbf Z ) & \overset{\psi _ n}{\simeq} &H ^ {n} ( G( L/K ), C _ {L} ) \\ \mathop{\rm cores} \uparrow &{} &\uparrow {f } \\ H ^ {n- 2} ( G( L _ {v} / K _ {v} ), \mathbf Z ) & \underset{\psi_{n}}{\simeq} &H ^ {n} ( G( L _ {v} / K _ {v} ), L _ {v} ^ {*} ) , \\ \end{array} $$

where the mapping $ f $ is induced by the imbedding $ L _ {v} ^ {*} \rightarrow I _ {L} \rightarrow C _ {L} $ and the co-restriction mapping cores. For $ n = 0 $, (3) gives the commutative diagram

$$ \tag{4 } \begin{array}{ccc} G( L/K )/[ G( L/K ), G( L/K )] & \overset{\psi}{\simeq} &C _ {K} /N _ {L/K} ( C _ {L} ) \\ \uparrow &{} &\uparrow \\ G( L _ {v} / K _ {v} )/[ G( L _ {v} /K _ {v} ), G( L _ {v} / K _ {v} )] & \underset{\phi}{\simeq} &K _ {v} ^ {*} / N _ {L _ {v} /K _ {v} } ( L _ {v} ^ {*} ). \\ \end{array} $$

The diagram (4) enables one to obtain a decomposition law of prime divisors of the field $ K $ in the Abelian extension $ L/K $. That is, a prime divisor $ \mathfrak c $ of $ K $ is unramified (splits completely) in $ L $ if and only if $ U( K _ {\mathfrak c } ) \subset N _ {L/K} ( C _ {L} ) $( correspondingly, $ K ^ {*} \subset N _ {L/K} ( C _ {L} ) $).

If $ \mathfrak c $ is a prime divisor of $ K $ that is unramified in $ L $, $ v $ is the valuation of $ K $ corresponding to $ \mathfrak c $ and $ \pi $ is a prime element of $ K _ {v} $, then the Artin symbol

$$ \left ( \frac{L/K}{\mathfrak c} \right ) = \psi ^ {- 1} ( \pi ) \in G( L/K ) $$

is defined and only depends on $ \mathfrak c $. It is the Frobenius automorphism in the decomposition subgroup of $ v $. According to Chebotarev's density theorem, any element of the group $ G( L/K ) $ has the form

$$ \left ( \frac{L/K}{\mathfrak c} \right ) $$

for an infinite number of prime divisors $ \mathfrak c $ of $ K $.

For example, the maximal unramified Abelian extension $ F $ of a number field $ K $ (called the Hilbert class field) is a field whose norm subgroup coincides with the image under the projection $ I _ {K} \rightarrow C _ {K} $ of the group $ K ^ {*} \prod _ {v} U( K _ {v} ) $, where $ v $ runs through all points of $ K $. The group $ I _ {K} /K ^ {*} \prod _ {v} U( K _ {v} ) $ is canonically isomorphic to the class group $ \mathop{\rm Cl} _ {K} $ of $ K $, which gives the important isomorphism $ G( F/K ) \simeq \mathop{\rm Cl} _ {K} $. In particular, there are no unramified Abelian extensions of $ K $ if and only if $ K $ has class number one.

The type of decomposition for a prime divisor $ \mathfrak c $ of the field $ K $ in $ F $ is completely determined by the class of $ \mathfrak c $ in $ \mathop{\rm Cl} _ {K} $. In particular, $ \mathfrak c $ splits completely if and only if $ \mathfrak c $ is principal. All divisors of $ K $ become principal divisors in $ F $.

Just as class field theory for unramified Abelian extensions can be explained in terms of the divisor class group and its subgroups, so can arbitrary Abelian extensions be characterized by means of ray class groups with respect to suitable modules (see Algebraic number theory). There are also generalizations of class field theory to the case of infinite Galois extensions [4].

Although class field theory arose as a theory on Abelian extensions, the results give important information also for non-Abelian Galois extensions. For example, class field theory is used in proving the existence of infinite class field towers (see Tower of fields).

References

[1] J.W.S. Cassels (ed.) A. Fröhlich (ed.) , Algebraic number theory , Acad. Press (1967) pp. Chapt. VI
[2] A. Weil, "Basic number theory" , Springer (1973)
[3] H. Koch, "Galoissche Theorie der $p$-Erweiterungen" , Deutsch. Verlag Wissenschaft. (1970)
[4] L.V. Kuz'min, "Homotopy of profinite groups, the Schur multiplicator and class field theory" Izv. Akad. Nauk SSSR Ser. Mat. , 33 : 6 (1969) pp. 1220–1254 (In Russian)

Comments

Let $ A $ be the ring of integers of the global field $ K $. Then the class group $ \mathop{\rm Cl} _ {K} $ is the divisor class group of $ A $; i.e. it is the group of classes of ideals of $ A $ modulo principal ideals.

The fact that the divisors of a field $ K $ become principal divisors in its maximal unramified Abelian extension $ F $ is called the principal ideal theorem.

Two excellent modern up-to-date books on class field theory are [a1] and [a2]. The latter also discusses the relations between the idèle-theoretic formulation of class field theory and ray class groups.

The Kronecker–Weber theorem states that every Abelian finite extension of $ \mathbf Q $ is contained in some $ \mathbf Q ( \xi _ {n} ) $ where $ \xi _ {n} $ is a primitive $ n $-th root of unity (i.e. $ \xi _ {n} ^ {n} = 1 $ and $ \xi _ {n} ^ {m} \neq 1 $ for $ m < n $). Kronecker also conjectured that every Abelian extension of an imaginary quadratic field $ \mathbf Q ( \sqrt{ - d} ) $, $ d \in \mathbf N $, is contained in an extension generated by the torsion points of an elliptic curve with complex multiplication. This was proved by T. Takagi [a3]. Its analogue for local fields is the Lubin–Tate theorem, stating that the torsion points of a Lubin–Tate formal group over the ring of integers $ A $ of a local field $ K $ together with the maximal unramified extension of $ K $ generate the maximal Abelian extension of $ K $[a4]. These formal groups can be used to give very explicit descriptions of the local reciprocity mappings $ K ^ {*} \rightarrow \mathop{\rm Gal} ( K ^ {ab} / K ) $, cf. also [a5]. The Lubin–Tate formal groups are analogous to elliptic curves with complex multiplication in that they have maximally large endomorphism rings.

The mapping $ L \rightarrow N _ {L/K} C _ {L} \subset C _ {K} $ sets up a one-to-one correspondence between the finite Abelian extensions $ L / K $ and the closed subgroups of finite index in the idèle class group $ C _ {K} $ of $ K $ (cf. Idèle for the topology on $ C _ {K} $). This is often called the existence theorem of class field theory. If $ L / K $ is associated to a subgroup $ N $ of $ C _ {K} $, then $ L / K $ is called the class field of $ N $. Let $ \mathfrak m = \prod _ {\mathfrak p} \mathfrak p ^ {n _ {\mathfrak p} } $ be a formal product of prime divisors of $ K $ such that $ n _ {\mathfrak p} \geq 0 $ for all $ \mathfrak p $, $ n _ {\mathfrak p} = 0 $ for almost-all $ \mathfrak p $ and $ n _ {\mathfrak p} = 0 $ or 1 for the infinite primes of $ K $. Such a formal product is called a positive divisor or cycle. For each prime divisor $ \mathfrak p $, let $ K _ {\mathfrak p} $ be the local field obtained by completion with respect to the valuation defined by $ \mathfrak p $ and let $ A _ {\mathfrak p} $ be its ring of integers. For each finite prime, let $ U _ {\mathfrak p} ^ {n} = \{ {x \in A _ {\mathfrak p} } : {x \equiv 1 \mathop{\rm mod} \mathfrak p ^ {n} } \} $ for $ n > 0 $ and $ U _ {\mathfrak p} ^ {0} = U _ {\mathfrak p} = A _ {\mathfrak p} ^ {*} $, the group of units of $ A _ {\mathfrak p} $. In addition, for the infinite primes $ \mathfrak p $ define $ U _ {\mathfrak p} ^ {1} = \mathbf R ^ {+} $, the positive reals, if $ \mathfrak p $ is real, $ U _ {\mathfrak p} ^ {0} = \mathbf R ^ {*} $ if $ \mathfrak p $ is real, and $ U _ {\mathfrak p} ^ {0} = U _ {\mathfrak p} ^ {1} = \mathbf C ^ {*} $ if $ \mathfrak p $ is complex. Given a positive divisor $ \mathfrak n $, a corresponding subgroup $ C _ {K} ^ {\mathfrak n} $ of $ C _ {K} $ is defined by $ C _ {k} ^ {\mathfrak n} = I _ {K} ^ {\mathfrak n} K ^ {*} / K ^ {*} $ where $ I _ {K} ^ {\mathfrak n} $ is the subgroup of the group of idèles, $ I _ {K} $, defined by

$$ I _ {K} ^ {\mathfrak n} = \{ {\alpha \in I _ {K} } : {\alpha _ {\mathfrak p} \in U _ {\mathfrak p} ^ {n _ {\mathfrak p} } \textrm{ for all } \mathfrak p } \} $$

The subgroup $ C _ {K} ^ {\mathfrak n} $ is called a congruence subgroup; more precisely, it is the congruence subgroup $ \mathop{\rm mod} \mathfrak n $ of $ C _ {K} $. The corresponding class field $ K ^ {\mathfrak n} / K $, i.e. the Abelian extension such that $ N _ {K ^ {\mathfrak n} / K } C _ {K ^ {\mathfrak n} } = C _ {K} ^ {\mathfrak n} $, is called the ray class field $ \mathop{\rm mod} \mathfrak n $. Of particular interest is the ray class field $ \mathop{\rm mod} 1 $, which is the Hilbert class field, since $ C _ {K} / C _ {K} ^ {1} $ is clearly isomorphic to the ideal class group of all ideals modulo principal ideals.

References

[a1] K. Iwasawa, "Local class field theory" , Oxford Univ. Press (1986)
[a2] J. Neukirch, "Class field theory" , Springer (1986) pp. Chapt. 4, Sect. 8
[a3] T. Takagi, "Ueber eine Theorie des relativ-abelschen Zahlkörpers" J. Coll. Sci. Imp. Univ. Tokyo , 41 (1920) pp. 1–132
[a4] J. Lubin, J. Tate, "Formal complex multiplication in local fields" Ann. of Math. , 81 (1965) pp. 380–387
[a5] M. Hazewinkel, "Local class field theory is easy" Adv. in Math. , 18 (1975) pp. 148–181
How to Cite This Entry:
Class field theory. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Class_field_theory&oldid=17596
This article was adapted from an original article by L.V. Kuz'min (originator), which appeared in Encyclopedia of Mathematics - ISBN 1402006098. See original article