Namespaces
Variants
Actions

Difference between revisions of "Artin root numbers"

From Encyclopedia of Mathematics
Jump to: navigation, search
(Importing text file)
 
m (AUTOMATIC EDIT (latexlist): Replaced 128 formulas out of 130 by TEX code with an average confidence of 2.0 and a minimal confidence of 2.0.)
Line 1: Line 1:
A global Artin root number is a [[Complex number|complex number]] <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a1202701.png" /> of modulus <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a1202702.png" /> appearing in the functional equation of an Artin <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a1202704.png" />-series (cf. also [[L-function|<img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a1202705.png" />-function]])
+
<!--This article has been texified automatically. Since there was no Nroff source code for this article,
 +
the semi-automatic procedure described at https://encyclopediaofmath.org/wiki/User:Maximilian_Janisch/latexlist
 +
was used.
 +
If the TeX and formula formatting is correct, please remove this message and the {{TEX|semi-auto}} category.
  
<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/a/a120/a120270/a1202706.png" /></td> <td valign="top" style="width:5%;text-align:right;">(a1)</td></tr></table>
+
Out of 130 formulas, 128 were replaced by TEX code.-->
  
in which <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a1202707.png" /> is a representation
+
{{TEX|semi-auto}}{{TEX|partial}}
 +
A global Artin root number is a [[Complex number|complex number]] $W ( \rho )$ of modulus $1$ appearing in the functional equation of an Artin $L$-series (cf. also [[L-function|$L$-function]])
  
<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/a/a120/a120270/a1202708.png" /></td> </tr></table>
+
\begin{equation} \tag{a1} \Lambda ( s , \rho ) = W ( \rho ) . \Lambda ( 1 - s , \overline { \rho } ), \end{equation}
  
of the [[Galois group|Galois group]] of a finite Galois extension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a1202709.png" /> of global fields (cf. also [[Representation theory|Representation theory]]; [[Galois theory|Galois theory]]; [[Extension of a field|Extension of a field]]), <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027010.png" /> denotes the complex-conjugate representation, and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027011.png" /> is the (extended) Artin <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027012.png" />-series with gamma factors at the Archimedean places of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027013.png" /> (details can be found in [[#References|[a6]]]).
+
in which $\rho$ is a representation
 +
 
 +
\begin{equation*} \rho : \operatorname { Gal } ( N / K ) \rightarrow G l _ { n } ( C ) \end{equation*}
 +
 
 +
of the [[Galois group|Galois group]] of a finite Galois extension $N / K$ of global fields (cf. also [[Representation theory|Representation theory]]; [[Galois theory|Galois theory]]; [[Extension of a field|Extension of a field]]), $\rho$ denotes the complex-conjugate representation, and $\Lambda ( s , \rho )$ is the (extended) Artin $L$-series with gamma factors at the Archimedean places of $K$ (details can be found in [[#References|[a6]]]).
  
 
Work of R. Langlands (unpublished) and P. Deligne [[#References|[a2]]] shows that the global Artin root number can be written canonically as a product
 
Work of R. Langlands (unpublished) and P. Deligne [[#References|[a2]]] shows that the global Artin root number can be written canonically as a product
  
<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/a/a120/a120270/a12027014.png" /></td> </tr></table>
+
\begin{equation*} W ( \rho ) = \prod W _ { P } ( \rho ) \end{equation*}
  
of other complex numbers of modulus <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027015.png" />, called local Artin root numbers (Deligne calls them simply  "local constants" ). Given <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027016.png" />, there is one local root number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027017.png" /> for each non-trivial place <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027018.png" /> of the base field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027019.png" />, and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027020.png" /> for almost all <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027021.png" />.
+
of other complex numbers of modulus $1$, called local Artin root numbers (Deligne calls them simply  "local constants" ). Given $\rho$, there is one local root number $W _ { P } ( \rho )$ for each non-trivial place $P$ of the base field $K$, and $W _ { P } ( \rho ) = 1$ for almost all $P$.
  
Interest in root numbers arises in part because they are analogues of Langlands' <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027023.png" />-factors appearing in the functional equations of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027024.png" />-series associated to automorphic forms. In special settings, global root numbers are known to have deep connections to the vanishing of Dedekind zeta-functions at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027025.png" /> (cf. also [[Dedekind zeta-function|Dedekind zeta-function]]), and to the existence of a global normal integral basis, while local root numbers are connected to Stiefel–Whitney classes, to Hasse symbols of trace forms, and to the existence of a canonical quadratic refinement of the local Hilbert symbol. Excellent references containing both a general account as well as details can be found in [[#References|[a6]]], [[#References|[a11]]] and [[#References|[a4]]].
+
Interest in root numbers arises in part because they are analogues of Langlands' <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027023.png"/>-factors appearing in the functional equations of $L$-series associated to automorphic forms. In special settings, global root numbers are known to have deep connections to the vanishing of Dedekind zeta-functions at $s = 1 / 2$ (cf. also [[Dedekind zeta-function|Dedekind zeta-function]]), and to the existence of a global normal integral basis, while local root numbers are connected to Stiefel–Whitney classes, to Hasse symbols of trace forms, and to the existence of a canonical quadratic refinement of the local Hilbert symbol. Excellent references containing both a general account as well as details can be found in [[#References|[a6]]], [[#References|[a11]]] and [[#References|[a4]]].
  
 
==Some observations.==
 
==Some observations.==
  
  
a) The global and the local root numbers of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027026.png" /> depend only on the isomorphism class of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027027.png" />; hence the root numbers are functions of the character <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027028.png" />.
+
a) The global and the local root numbers of $\rho$ depend only on the isomorphism class of $\rho$; hence the root numbers are functions of the character $\chi = \text { trace o } \rho$.
  
b) When the character of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027029.png" /> is real-valued, then the global root number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027030.png" /> has value <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027031.png" />, and each local root number is a fourth root of unity.
+
b) When the character of $\rho$ is real-valued, then the global root number $W ( \rho )$ has value $\pm 1$, and each local root number is a fourth root of unity.
  
c) [[#References|[a1]]] When <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027032.png" /> has a representation <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027033.png" /> whose character is real-valued and whose global root number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027034.png" /> is <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027035.png" />, then the Dedekind zeta-function <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027036.png" /> vanishes at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027037.png" />.
+
c) [[#References|[a1]]] When $\operatorname{Gal}( N / K )$ has a representation $\rho$ whose character is real-valued and whose global root number $W ( \rho )$ is $- 1$, then the Dedekind zeta-function $\zeta_N ( s )$ vanishes at $s = 1 / 2$.
  
d) [[#References|[a5]]] When <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027038.png" /> is a real representation (a condition stronger than the requirement that the character be real-valued), then the global root number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027039.png" /> is <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027040.png" />. This means that the product of the local root numbers of a real representation is <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027041.png" />, so the Fröhlich–Queyrut theorem is a reciprocity law (cf. [[Reciprocity laws|Reciprocity laws]]), or a  "product formula" , for local root numbers. Some authors write  "real orthogonal"  or just  "orthogonal"  in place of  "real representation" ; all three concepts are equivalent.
+
d) [[#References|[a5]]] When $\rho$ is a real representation (a condition stronger than the requirement that the character be real-valued), then the global root number $W ( \rho )$ is $+ 1$. This means that the product of the local root numbers of a real representation is $+ 1$, so the Fröhlich–Queyrut theorem is a reciprocity law (cf. [[Reciprocity laws|Reciprocity laws]]), or a  "product formula" , for local root numbers. Some authors write  "real orthogonal"  or just  "orthogonal"  in place of  "real representation" ; all three concepts are equivalent.
  
e) [[#References|[a12]]] A normal extension of number fields <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027042.png" /> has a normal integral basis if and only if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027043.png" /> is at most tamely ramified and the global root number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027044.png" /> for all irreducible symplectic representations <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027045.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027046.png" />. (By definition, the extension has a normal integral basis provided the ring of integers <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027047.png" /> is a free <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027048.png" />-module).
+
e) [[#References|[a12]]] A normal extension of number fields $N / K$ has a normal integral basis if and only if $N / K$ is at most tamely ramified and the global root number $W ( \rho ) = 1$ for all irreducible symplectic representations $\rho$ of $\operatorname{Gal}( N / K )$. (By definition, the extension has a normal integral basis provided the ring of integers $O _ { \text{N} }$ is a free $\mathbf{Z}[ \text{Gal} (N/K)]$-module).
  
f) [[#References|[a3]]] Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027049.png" /> be a real representation and let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027050.png" /> be the <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027051.png" />-dimensional real representation obtained by composing <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027052.png" /> with the determinant. Then for each place <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027053.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027054.png" />, the normalized local Artin root number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027055.png" /> equals the second [[Stiefel–Whitney class|Stiefel–Whitney class]] <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027056.png" /> of the restriction of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027057.png" /> to a decomposition subgroup of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027058.png" /> in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027059.png" />.
+
f) [[#References|[a3]]] Let $\rho$ be a real representation and let $\operatorname{det}_ { \rho }$ be the $1$-dimensional real representation obtained by composing $\rho$ with the determinant. Then for each place $P$ of $K$, the normalized local Artin root number $W _ { P } ( \rho ) / W _ { P } ( \operatorname { det } _ { \rho } )$ equals the second [[Stiefel–Whitney class|Stiefel–Whitney class]] $w _ { 2 } ( \rho _ { P } )$ of the restriction of $\rho$ to a decomposition subgroup of $P$ in $\operatorname{Gal}( N / K )$.
  
g) [[#References|[a9]]] Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027060.png" /> be a finite extension of number fields, with normal closure <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027061.png" />. Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027062.png" /> be the representation of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027063.png" /> induced by the trivial representation of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027064.png" />. Then the Hasse symbol <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027065.png" /> at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027066.png" /> of the trace form <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027067.png" /> is given by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027068.png" />, where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027069.png" /> is the discriminant of the trace form and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027070.png" /> is a Hilbert symbol (cf. [[Norm-residue symbol|Norm-residue symbol]]).
+
g) [[#References|[a9]]] Let $E / K$ be a finite extension of number fields, with normal closure $N / K$. Let $\rho$ be the representation of $\operatorname{Gal}( N / K )$ induced by the trivial representation of $\operatorname { Gal } ( N / E )$. Then the Hasse symbol $h _ { P }$ at $P$ of the trace form $\operatorname {trace}_{E/K} ( x ^ { 2 } )$ is given by $h _ { p } = ( 2 , d ) _ { P } \cdot W _ { P } ( \rho ) / W _ { P } ( \operatorname { det } _ { \rho } )$, where $d$ is the discriminant of the trace form and $( 2 , d ) _ { P }$ is a Hilbert symbol (cf. [[Norm-residue symbol|Norm-residue symbol]]).
  
h) [[#References|[a11]]] For a place <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027071.png" /> and a non-zero element <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027072.png" />, the <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027073.png" />-dimensional real representation <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027074.png" /> sending <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027075.png" /> to <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027076.png" /> has a local root number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027077.png" />, which will be abbreviated by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027078.png" />. For <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027079.png" /> fixed, these local root numbers produce a mapping
+
h) [[#References|[a11]]] For a place $P$ and a non-zero element $a \in K ^ { * }$, the $1$-dimensional real representation $\rho _ { a }$ sending $g \in G$ to $\rho _ { a } ( g ) = g ( \sqrt { a } ) / \sqrt { a }$ has a local root number $W _ { P } (\, \rho _ { a } )$, which will be abbreviated by $r _ { P } ( a )$. For $P$ fixed, these local root numbers produce a mapping
  
<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/a/a120/a120270/a12027080.png" /></td> </tr></table>
+
\begin{equation*} r _ { P } : K _ { P } ^ { * } / K _ { P } ^ { * 2 } \rightarrow C ^ { * } \end{equation*}
  
 
which satisfies
 
which satisfies
  
<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/a/a120/a120270/a12027081.png" /></td> <td valign="top" style="width:5%;text-align:right;">(a2)</td></tr></table>
+
\begin{equation} \tag{a2} r _ { P } ( a \cdot b ) = r _ { P } ( a ) \cdot r _ { P } ( b ) \cdot ( a , b ) _ { P }. \end{equation}
  
The last factor is the local Hilbert symbol at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027082.png" />; it gives a non-degenerate [[Inner product|inner product]] on the local square class group at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027083.png" />, viewed as a vector space over the field of two elements, the latter identified with <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027084.png" />. Equation (a2) has been interpreted in [[#References|[a7]]] to mean that the local root numbers give a canonical  "quadratic refinement"  of this inner product.
+
The last factor is the local Hilbert symbol at $P$; it gives a non-degenerate [[Inner product|inner product]] on the local square class group at $P$, viewed as a vector space over the field of two elements, the latter identified with $\{ \pm 1 \}$. Equation (a2) has been interpreted in [[#References|[a7]]] to mean that the local root numbers give a canonical  "quadratic refinement"  of this inner product.
  
 
==Remarks.==
 
==Remarks.==
  
  
1) It follows formally from (a1) that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027085.png" />. Moreover, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027086.png" />, so the global root number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027087.png" /> has modulus <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027088.png" />. When the character of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027089.png" /> is real-valued, then <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027090.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027091.png" /> are isomorphic, so their global root numbers are equal: <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027092.png" />. It follows that the global root number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027093.png" />.
+
1) It follows formally from (a1) that $W ( \rho ) \cdot W ( \overline { \rho } ) = 1$. Moreover, $W ( \overline { \rho } ) = \overline { W ( \rho ) }$, so the global root number $W ( \rho )$ has modulus $1$. When the character of $\rho$ is real-valued, then $\rho$ and $\rho$ are isomorphic, so their global root numbers are equal: $W ( \rho ) = W ( \overline { \rho } )$. It follows that the global root number $W ( \rho ) = \pm 1$.
  
 
2) Statement c) follows from the basic argument in [[#References|[a1]]], Sect. 3, with minor modifications.
 
2) Statement c) follows from the basic argument in [[#References|[a1]]], Sect. 3, with minor modifications.
  
3) To put Taylor's theorem in context, let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027094.png" /> be a finite Galois extension of fields, with Galois group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027095.png" />. Then the normal basis theorem of field theory says that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027096.png" /> has a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027097.png" />-basis consisting of the Galois conjugates of a single element; restated, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027098.png" /> is a free <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a12027099.png" />-module. When <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270100.png" /> is an extension of number fields, one can ask for a normal integral basis. There are two different notions: One can require the ring of integers <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270101.png" /> to be a free <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270102.png" /> module (necessarily of rank <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270103.png" />), or one can require <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270104.png" /> to be a free <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270105.png" />-module (necessarily of rank <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270106.png" />). These notions coincide when the base field <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270107.png" /> is the field of rational numbers. At present (1998), little is known about the first notion, so the second is chosen. Thus, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270108.png" /> has a normal integral basis when <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270109.png" /> has a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270110.png" />-basis <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270111.png" />. By results of E. Noether and R. Swan (see [[#References|[a4]]], pp. 26–28), a necessary condition for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270112.png" /> to have a normal integral basis is that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270113.png" /> be at most tamely ramified. A. Fröhlich conjectured and M. Taylor proved that the extra conditions beyond tameness needed to make <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270114.png" /> a free <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270115.png" />-module is for all the global root numbers of symplectic representations to have value <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270116.png" />.
+
3) To put Taylor's theorem in context, let $N / K$ be a finite Galois extension of fields, with Galois group $G$. Then the normal basis theorem of field theory says that $N$ has a $K$-basis consisting of the Galois conjugates of a single element; restated, $N$ is a free $K [ G ]$-module. When $N / K$ is an extension of number fields, one can ask for a normal integral basis. There are two different notions: One can require the ring of integers $O _ { \text{N} }$ to be a free $O _ { K } [G]$ module (necessarily of rank $1$), or one can require $O _ { \text{N} }$ to be a free $\mathbf Z [ G ]$-module (necessarily of rank $[ K : Q ]$). These notions coincide when the base field $K$ is the field of rational numbers. At present (1998), little is known about the first notion, so the second is chosen. Thus, $N / K$ has a normal integral basis when $O _ { \text{N} }$ has a $\mathbf{Z}$-basis $\{ a _ { j } ^ { g } : j = 1 , \dots , [ K : Q ] , g \in G \}$. By results of E. Noether and R. Swan (see [[#References|[a4]]], pp. 26–28), a necessary condition for $N / K$ to have a normal integral basis is that $N / K$ be at most tamely ramified. A. Fröhlich conjectured and M. Taylor proved that the extra conditions beyond tameness needed to make $O _ { \text{N} }$ a free $\mathbf Z [ G ]$-module is for all the global root numbers of symplectic representations to have value $1$.
  
To say that a complex representation <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270117.png" /> is symplectic means that the representation has even dimension, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270118.png" />, and factors through the [[Symplectic group|symplectic group]] <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270119.png" />. The character values of a symplectic representation are real. A useful criterion is: When <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270120.png" /> is irreducible with character <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270121.png" />, then the sum
+
To say that a complex representation $\rho$ is symplectic means that the representation has even dimension, $2 n$, and factors through the [[Symplectic group|symplectic group]] $\rho : G \rightarrow S p _ { 2 n } ( C ) \rightarrow G l _ { 2 n } ( C )$. The character values of a symplectic representation are real. A useful criterion is: When $\rho$ is irreducible with character <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270121.png"/>, then the sum
  
<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/a/a120/a120270/a120270122.png" /></td> </tr></table>
+
\begin{equation*} | G | ^ { - 1 } \sum _ { g \in G } \chi ( g ^ { 2 } ) \end{equation*}
  
takes the value <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270123.png" /> when <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270124.png" /> is symplectic, the value <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270125.png" /> when <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270126.png" /> is real, and the value <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270127.png" /> in all other cases (see [[#References|[a10]]], Prop. 39).
+
takes the value $- 1$ when $\rho$ is symplectic, the value $1$ when $\rho$ is real, and the value $0$ in all other cases (see [[#References|[a10]]], Prop. 39).
  
4) The families of complex numbers which can be realized as the local root numbers of some real representation of the Galois group of some normal extension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270128.png" /> have been determined in [[#References|[a8]]].
+
4) The families of complex numbers which can be realized as the local root numbers of some real representation of the Galois group of some normal extension $N / Q$ have been determined in [[#References|[a8]]].
  
 
====References====
 
====References====
<table><TR><TD valign="top">[a1]</TD> <TD valign="top">  J.V. Armitage,  "Zeta functions with zero at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270129.png" />"  ''Invent. Math.'' , '''15'''  (1972)  pp. 199–205</TD></TR><TR><TD valign="top">[a2]</TD> <TD valign="top">  P. Deligne,  "Les constantes des équation fonctionelles des fonctions <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270130.png" />" , ''Lecture Notes Math.'' , '''349''' , Springer  (1974)  pp. 501–597</TD></TR><TR><TD valign="top">[a3]</TD> <TD valign="top">  P. Deligne,  "Les constantes locales de l'équation fonctionelle des fonction <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270131.png" /> d'Artin d'une répresentation orthogonale"  ''Invent. Math.'' , '''35'''  (1976)  pp. 299–316</TD></TR><TR><TD valign="top">[a4]</TD> <TD valign="top">  A. Fröhlich,  "Galois module structure of algebraic integers" , ''Ergebn. Math.'' , '''1''' , Springer  (1983)</TD></TR><TR><TD valign="top">[a5]</TD> <TD valign="top">  A. Fröhlich,  J. Queyrut,  "On the functional equation of the Artin L-function for characters of real representations"  ''Invent. Math.'' , '''20'''  (1973)  pp. 125–138</TD></TR><TR><TD valign="top">[a6]</TD> <TD valign="top">  J. Martinet,  "Character theory and Artin L-functions" , ''Algebraic Number Fields: Proc. Durham Symp. 1975'' , Acad. Press  (1977)  pp. 1–87</TD></TR><TR><TD valign="top">[a7]</TD> <TD valign="top">  R. Perlis,  "Arf equivalence I" , ''Number Theory in Progress: Proc. Internat. Conf. in Honor of A. Schinzel (Zakopane, Poland, June 30--July 9, 1997)'' , W. de Gruyter  (1999)</TD></TR><TR><TD valign="top">[a8]</TD> <TD valign="top">  R. Perlis,  "On the analytic determination of the trace form"  ''Canad. Math. Bull.'' , '''28''' :  4  (1985)  pp. 422–430</TD></TR><TR><TD valign="top">[a9]</TD> <TD valign="top">  J-P. Serre,  "L'invariant de Witt de la forme <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120270/a120270132.png" />"  ''Comment. Math. Helvetici'' , '''59'''  (1984)  pp. 651–676</TD></TR><TR><TD valign="top">[a10]</TD> <TD valign="top">  J-P. Serre,  "Représentations linéaires des groupes finis" , Hermann  (1971)  (Edition: Second)</TD></TR><TR><TD valign="top">[a11]</TD> <TD valign="top">  J. Tate,  "Local constants" , ''Algebraic Number Fields: Proc. Durham Symp. 1975'' , Acad. Press  (1977)  pp. 89–131</TD></TR><TR><TD valign="top">[a12]</TD> <TD valign="top">  M. Taylor,  "On Fröhlich's conjecture for rings of integers of tame extensions"  ''Invent. Math.'' , '''63'''  (1981)  pp. 41–79</TD></TR></table>
+
<table><tr><td valign="top">[a1]</td> <td valign="top">  J.V. Armitage,  "Zeta functions with zero at $s = 1 / 2$"  ''Invent. Math.'' , '''15'''  (1972)  pp. 199–205</td></tr><tr><td valign="top">[a2]</td> <td valign="top">  P. Deligne,  "Les constantes des équation fonctionelles des fonctions $L$" , ''Lecture Notes Math.'' , '''349''' , Springer  (1974)  pp. 501–597</td></tr><tr><td valign="top">[a3]</td> <td valign="top">  P. Deligne,  "Les constantes locales de l'équation fonctionelle des fonction $L$ d'Artin d'une répresentation orthogonale"  ''Invent. Math.'' , '''35'''  (1976)  pp. 299–316</td></tr><tr><td valign="top">[a4]</td> <td valign="top">  A. Fröhlich,  "Galois module structure of algebraic integers" , ''Ergebn. Math.'' , '''1''' , Springer  (1983)</td></tr><tr><td valign="top">[a5]</td> <td valign="top">  A. Fröhlich,  J. Queyrut,  "On the functional equation of the Artin L-function for characters of real representations"  ''Invent. Math.'' , '''20'''  (1973)  pp. 125–138</td></tr><tr><td valign="top">[a6]</td> <td valign="top">  J. Martinet,  "Character theory and Artin L-functions" , ''Algebraic Number Fields: Proc. Durham Symp. 1975'' , Acad. Press  (1977)  pp. 1–87</td></tr><tr><td valign="top">[a7]</td> <td valign="top">  R. Perlis,  "Arf equivalence I" , ''Number Theory in Progress: Proc. Internat. Conf. in Honor of A. Schinzel (Zakopane, Poland, June 30--July 9, 1997)'' , W. de Gruyter  (1999)</td></tr><tr><td valign="top">[a8]</td> <td valign="top">  R. Perlis,  "On the analytic determination of the trace form"  ''Canad. Math. Bull.'' , '''28''' :  4  (1985)  pp. 422–430</td></tr><tr><td valign="top">[a9]</td> <td valign="top">  J-P. Serre,  "L'invariant de Witt de la forme $\operatorname { Tr } ( x ^ { 2 } )$"  ''Comment. Math. Helvetici'' , '''59'''  (1984)  pp. 651–676</td></tr><tr><td valign="top">[a10]</td> <td valign="top">  J-P. Serre,  "Représentations linéaires des groupes finis" , Hermann  (1971)  (Edition: Second)</td></tr><tr><td valign="top">[a11]</td> <td valign="top">  J. Tate,  "Local constants" , ''Algebraic Number Fields: Proc. Durham Symp. 1975'' , Acad. Press  (1977)  pp. 89–131</td></tr><tr><td valign="top">[a12]</td> <td valign="top">  M. Taylor,  "On Fröhlich's conjecture for rings of integers of tame extensions"  ''Invent. Math.'' , '''63'''  (1981)  pp. 41–79</td></tr></table>

Revision as of 16:59, 1 July 2020

A global Artin root number is a complex number $W ( \rho )$ of modulus $1$ appearing in the functional equation of an Artin $L$-series (cf. also $L$-function)

\begin{equation} \tag{a1} \Lambda ( s , \rho ) = W ( \rho ) . \Lambda ( 1 - s , \overline { \rho } ), \end{equation}

in which $\rho$ is a representation

\begin{equation*} \rho : \operatorname { Gal } ( N / K ) \rightarrow G l _ { n } ( C ) \end{equation*}

of the Galois group of a finite Galois extension $N / K$ of global fields (cf. also Representation theory; Galois theory; Extension of a field), $\rho$ denotes the complex-conjugate representation, and $\Lambda ( s , \rho )$ is the (extended) Artin $L$-series with gamma factors at the Archimedean places of $K$ (details can be found in [a6]).

Work of R. Langlands (unpublished) and P. Deligne [a2] shows that the global Artin root number can be written canonically as a product

\begin{equation*} W ( \rho ) = \prod W _ { P } ( \rho ) \end{equation*}

of other complex numbers of modulus $1$, called local Artin root numbers (Deligne calls them simply "local constants" ). Given $\rho$, there is one local root number $W _ { P } ( \rho )$ for each non-trivial place $P$ of the base field $K$, and $W _ { P } ( \rho ) = 1$ for almost all $P$.

Interest in root numbers arises in part because they are analogues of Langlands' -factors appearing in the functional equations of $L$-series associated to automorphic forms. In special settings, global root numbers are known to have deep connections to the vanishing of Dedekind zeta-functions at $s = 1 / 2$ (cf. also Dedekind zeta-function), and to the existence of a global normal integral basis, while local root numbers are connected to Stiefel–Whitney classes, to Hasse symbols of trace forms, and to the existence of a canonical quadratic refinement of the local Hilbert symbol. Excellent references containing both a general account as well as details can be found in [a6], [a11] and [a4].

Some observations.

a) The global and the local root numbers of $\rho$ depend only on the isomorphism class of $\rho$; hence the root numbers are functions of the character $\chi = \text { trace o } \rho$.

b) When the character of $\rho$ is real-valued, then the global root number $W ( \rho )$ has value $\pm 1$, and each local root number is a fourth root of unity.

c) [a1] When $\operatorname{Gal}( N / K )$ has a representation $\rho$ whose character is real-valued and whose global root number $W ( \rho )$ is $- 1$, then the Dedekind zeta-function $\zeta_N ( s )$ vanishes at $s = 1 / 2$.

d) [a5] When $\rho$ is a real representation (a condition stronger than the requirement that the character be real-valued), then the global root number $W ( \rho )$ is $+ 1$. This means that the product of the local root numbers of a real representation is $+ 1$, so the Fröhlich–Queyrut theorem is a reciprocity law (cf. Reciprocity laws), or a "product formula" , for local root numbers. Some authors write "real orthogonal" or just "orthogonal" in place of "real representation" ; all three concepts are equivalent.

e) [a12] A normal extension of number fields $N / K$ has a normal integral basis if and only if $N / K$ is at most tamely ramified and the global root number $W ( \rho ) = 1$ for all irreducible symplectic representations $\rho$ of $\operatorname{Gal}( N / K )$. (By definition, the extension has a normal integral basis provided the ring of integers $O _ { \text{N} }$ is a free $\mathbf{Z}[ \text{Gal} (N/K)]$-module).

f) [a3] Let $\rho$ be a real representation and let $\operatorname{det}_ { \rho }$ be the $1$-dimensional real representation obtained by composing $\rho$ with the determinant. Then for each place $P$ of $K$, the normalized local Artin root number $W _ { P } ( \rho ) / W _ { P } ( \operatorname { det } _ { \rho } )$ equals the second Stiefel–Whitney class $w _ { 2 } ( \rho _ { P } )$ of the restriction of $\rho$ to a decomposition subgroup of $P$ in $\operatorname{Gal}( N / K )$.

g) [a9] Let $E / K$ be a finite extension of number fields, with normal closure $N / K$. Let $\rho$ be the representation of $\operatorname{Gal}( N / K )$ induced by the trivial representation of $\operatorname { Gal } ( N / E )$. Then the Hasse symbol $h _ { P }$ at $P$ of the trace form $\operatorname {trace}_{E/K} ( x ^ { 2 } )$ is given by $h _ { p } = ( 2 , d ) _ { P } \cdot W _ { P } ( \rho ) / W _ { P } ( \operatorname { det } _ { \rho } )$, where $d$ is the discriminant of the trace form and $( 2 , d ) _ { P }$ is a Hilbert symbol (cf. Norm-residue symbol).

h) [a11] For a place $P$ and a non-zero element $a \in K ^ { * }$, the $1$-dimensional real representation $\rho _ { a }$ sending $g \in G$ to $\rho _ { a } ( g ) = g ( \sqrt { a } ) / \sqrt { a }$ has a local root number $W _ { P } (\, \rho _ { a } )$, which will be abbreviated by $r _ { P } ( a )$. For $P$ fixed, these local root numbers produce a mapping

\begin{equation*} r _ { P } : K _ { P } ^ { * } / K _ { P } ^ { * 2 } \rightarrow C ^ { * } \end{equation*}

which satisfies

\begin{equation} \tag{a2} r _ { P } ( a \cdot b ) = r _ { P } ( a ) \cdot r _ { P } ( b ) \cdot ( a , b ) _ { P }. \end{equation}

The last factor is the local Hilbert symbol at $P$; it gives a non-degenerate inner product on the local square class group at $P$, viewed as a vector space over the field of two elements, the latter identified with $\{ \pm 1 \}$. Equation (a2) has been interpreted in [a7] to mean that the local root numbers give a canonical "quadratic refinement" of this inner product.

Remarks.

1) It follows formally from (a1) that $W ( \rho ) \cdot W ( \overline { \rho } ) = 1$. Moreover, $W ( \overline { \rho } ) = \overline { W ( \rho ) }$, so the global root number $W ( \rho )$ has modulus $1$. When the character of $\rho$ is real-valued, then $\rho$ and $\rho$ are isomorphic, so their global root numbers are equal: $W ( \rho ) = W ( \overline { \rho } )$. It follows that the global root number $W ( \rho ) = \pm 1$.

2) Statement c) follows from the basic argument in [a1], Sect. 3, with minor modifications.

3) To put Taylor's theorem in context, let $N / K$ be a finite Galois extension of fields, with Galois group $G$. Then the normal basis theorem of field theory says that $N$ has a $K$-basis consisting of the Galois conjugates of a single element; restated, $N$ is a free $K [ G ]$-module. When $N / K$ is an extension of number fields, one can ask for a normal integral basis. There are two different notions: One can require the ring of integers $O _ { \text{N} }$ to be a free $O _ { K } [G]$ module (necessarily of rank $1$), or one can require $O _ { \text{N} }$ to be a free $\mathbf Z [ G ]$-module (necessarily of rank $[ K : Q ]$). These notions coincide when the base field $K$ is the field of rational numbers. At present (1998), little is known about the first notion, so the second is chosen. Thus, $N / K$ has a normal integral basis when $O _ { \text{N} }$ has a $\mathbf{Z}$-basis $\{ a _ { j } ^ { g } : j = 1 , \dots , [ K : Q ] , g \in G \}$. By results of E. Noether and R. Swan (see [a4], pp. 26–28), a necessary condition for $N / K$ to have a normal integral basis is that $N / K$ be at most tamely ramified. A. Fröhlich conjectured and M. Taylor proved that the extra conditions beyond tameness needed to make $O _ { \text{N} }$ a free $\mathbf Z [ G ]$-module is for all the global root numbers of symplectic representations to have value $1$.

To say that a complex representation $\rho$ is symplectic means that the representation has even dimension, $2 n$, and factors through the symplectic group $\rho : G \rightarrow S p _ { 2 n } ( C ) \rightarrow G l _ { 2 n } ( C )$. The character values of a symplectic representation are real. A useful criterion is: When $\rho$ is irreducible with character , then the sum

\begin{equation*} | G | ^ { - 1 } \sum _ { g \in G } \chi ( g ^ { 2 } ) \end{equation*}

takes the value $- 1$ when $\rho$ is symplectic, the value $1$ when $\rho$ is real, and the value $0$ in all other cases (see [a10], Prop. 39).

4) The families of complex numbers which can be realized as the local root numbers of some real representation of the Galois group of some normal extension $N / Q$ have been determined in [a8].

References

[a1] J.V. Armitage, "Zeta functions with zero at $s = 1 / 2$" Invent. Math. , 15 (1972) pp. 199–205
[a2] P. Deligne, "Les constantes des équation fonctionelles des fonctions $L$" , Lecture Notes Math. , 349 , Springer (1974) pp. 501–597
[a3] P. Deligne, "Les constantes locales de l'équation fonctionelle des fonction $L$ d'Artin d'une répresentation orthogonale" Invent. Math. , 35 (1976) pp. 299–316
[a4] A. Fröhlich, "Galois module structure of algebraic integers" , Ergebn. Math. , 1 , Springer (1983)
[a5] A. Fröhlich, J. Queyrut, "On the functional equation of the Artin L-function for characters of real representations" Invent. Math. , 20 (1973) pp. 125–138
[a6] J. Martinet, "Character theory and Artin L-functions" , Algebraic Number Fields: Proc. Durham Symp. 1975 , Acad. Press (1977) pp. 1–87
[a7] R. Perlis, "Arf equivalence I" , Number Theory in Progress: Proc. Internat. Conf. in Honor of A. Schinzel (Zakopane, Poland, June 30--July 9, 1997) , W. de Gruyter (1999)
[a8] R. Perlis, "On the analytic determination of the trace form" Canad. Math. Bull. , 28 : 4 (1985) pp. 422–430
[a9] J-P. Serre, "L'invariant de Witt de la forme $\operatorname { Tr } ( x ^ { 2 } )$" Comment. Math. Helvetici , 59 (1984) pp. 651–676
[a10] J-P. Serre, "Représentations linéaires des groupes finis" , Hermann (1971) (Edition: Second)
[a11] J. Tate, "Local constants" , Algebraic Number Fields: Proc. Durham Symp. 1975 , Acad. Press (1977) pp. 89–131
[a12] M. Taylor, "On Fröhlich's conjecture for rings of integers of tame extensions" Invent. Math. , 63 (1981) pp. 41–79
How to Cite This Entry:
Artin root numbers. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Artin_root_numbers&oldid=50324
This article was adapted from an original article by R. Perlis (originator), which appeared in Encyclopedia of Mathematics - ISBN 1402006098. See original article