Page 338 -
P. 338

§3. Hasse–Weil L-Function and the Functional Equation  313

                                2
         (1) f v (T ) = 1 − a v T + q v T , where a v = q v + 1 − N v ,if E has good reduction at
            v.
         (2) f v (T ) = 1 − e v T , where

                        
                        −1If E has split multiplicative reduction at v,
                        
                    e v =  +1if E has nonsplit multiplicative reduction at v,
                        
                           0if E has additive reduction at v.
                        
        Observe that in all cases

                                             1
                                  N v = q v f v  .
                                             q v
        The complete non-Archimedean part of the L-function is    v  1/L v (s), where L v (s) =
         f v (q −s ).
            v
        (3.2) Definition. Let E/K be an elliptic curve over a number field K. The Hasse–
        Weil L-function of E over K is
                                               1
                                 L E/K (s) =      ,
                                             L v (s)
                                           v
        where the local factor L v (s) is given in the previous section (3.1). Other notations
        for L E/K (s) in current use are L E (s), L(s, E/K),and L(E/K, s).
           The question of the local factors at infinity is taken up by J.-P. Serre in Fac-
        teurs locaux des Fonctions Zˆ eta des Vari´ et´ es alg´ ebriques: S´ eminaire Delange-Pisot-
        Poitou, 11 mai 1970. Serre in this 1970 DPP seminar gives a general definition for
        any smooth projective variety over a global field.

        (3.3) Definition. Let E/K be an elliptic curve over a number field K. The modified
        Hasse–Weil L-function of E over K is

                            # E/K (s) = A s/2   K (s)L E/K (s),

        where A = A E/K and   K (s) are defined as follows:
         (1) The constant A E/K is given by

                               A E/K = N K/Q ( f E/K ) · d 2  ,
                                                    K/Q
            where d K/Q is the absolute discriminant of K and f E/K is the conductor of E
            over K, see 14(2.4) and Serre, DPP, pp. 19–12.
         (2) The gamma factor for the field K where n = [K : Q]is
                                                    n
                                              −s
                                   K (s) = [(2π)  (s)] .
   333   334   335   336   337   338   339   340   341   342   343