2. MARCO TEÓRICO
2.5. Fundamentación histórica
2.5.2. El deporte, su historia y la aparición de la mujer
ofcBF[m,aF,G]under projection to the module HIw1 (Q(μm)⊗Qp,∞,F−−DV1×V2(F⊗G)∗)is zero.
Proof By taking the Vi sufficiently small, we may assume that F−−DV1×V2(F ⊗G)∗
is actually isomorphic to RA(α−1), whereα = αFαG andA = O(V1×V2), and that
α−1<p1+handα−1 is not a zero-divisor. It suffices, therefore, to show thatL RA(α−1)
maps the image ofcBF[m,aF,G]to zero.
However, for each pair of integers (,)∈V1×V2with, 1+2hand such that F andG are not twists of each other, we know that the image ofLRA(α−1)(cBF[m,aF,G])
vanishes when restricted to (,)×W⊆Max(A)×W, by Proposition3.5.11. Since such pairs (,) are Zariski-dense in Max(A), the result follows.
Remark 7.1.3 Cf. [19, Lemma 8.1.5], which is an analogous (but rather stronger) statement in the ordinary case.
Hence the projection ofcBF[m,aF,G]toF−◦is in the image of the injection HIw1(Qp,∞,F−+DV1×V2(F⊗G)∗)→HIw1(Qp,∞,F−◦DV1×V2(F⊗G)∗).
SinceF+DV2(G)∗is isomorphic to an unramified module twisted by anA×-valued char-
acter of the cyclotomic Galois groupΓ, we may define a Perrin-Riou logarithm map for
F−+DV1×V2(F⊗G)∗by reparametrising the corresponding map for its unramified twist,
exactly as in Theorem 8.2.8 of [19]. That is, if we define D(F−+M(F⊗G)∗)=F−+D(F⊗G)∗(−1−κV2)
Γ=1
,
which is free of rank 1 overO(V1×V2), then we obtain the following theorem: Theorem 7.1.4 There is an injective morphism ofO(V1×V2×W)-modules
L:HIw1(Qp,∞,F−+DV1×V2(F⊗G)∗)→D(F−+M(F⊗G)∗) ˆ⊗O(W),
with the following property: for all classical specialisations f, g ofF,G, and all characters ofΓ of the formτ =j+ηwithηof finite order and j∈Z, we have a commutative diagram
HIw1 Qp,∞,F−+DV1×V2(F⊗G)∗ -L D(F−+ M(F⊗G)∗) ˆ⊗O(W) H1(Qp,F−+D(f ⊗g)∗(−j−η)) ? - F−+Dcris(M(f?⊗g)∗(−εg,p))
in which the bottom horizontal map is given by ⎧ ⎪ ⎨ ⎪ ⎩ 1− αpj fβg 1− αfβg p1+j −1 if r=0 p1+j αfβg r G(ε)−1 if r>0 ⎫ ⎪ ⎬ ⎪ ⎭· ⎧ ⎨ ⎩ (−1)k−j (k−j)! log if jk, (j−k−1)! exp∗ if j>k,
where exp∗ and log are the Bloch–Kato dual-exponential and logarithm maps, ε is the finite-order character εg,p ·η−1 of Γ, r 0 is the conductor of ε, and G(ε) =
a∈(Z/prZ)×ε(a)ζparis the Gauss sum.
Proof The construction of the mapL is immediate from (6.2.1). The content of the theorem is that the mapLrecovers the maps exp∗and log for the specialisations ofFand G; this follows from Nakamura’s construction of exp∗and log for (ϕ,Γ)-modules. Theorem 7.1.5 (Explicit reciprocity law)If the Viare sufficiently small, then we have
"
LcBF[1F,1,G]
,ηF⊗ωG#
=(c2−c−(k+k−2j)εF(c)−1εG(c)−1)(−1)1+jλN(F)−1Lp(F,G,1+j).
Here, Lp(F,G,1+j)denotes Urban’s3-variable p-adic L-function as constructed in[32], andεF andεG are the characters by which the prime-to-p diamond operators act onF andG.
Proof The two sides of the desired formula agree at every (k, k, j) withk ∈ V1,k ∈V2 and 0jmin(k, k), by [18, Theorem 6.5.9]. These points are manifestly Zariski-dense,
and the result follows.
Remark 7.1.6 The construction ofωGand the proof of the explicit reciprocity law are also valid ifGis a Coleman family passing through ap-stabilisationgαof ap-regular weight 1 form, as in Theorem4.7.2; the only difference is that one may need to replaceV2with a finite flat covering ˜V2. In this setting,gα is automatically ordinary, soGis in fact a Hida family, and one can use the construction ofωGgiven in [19, Proposition 10.12.2]. 8 Bounding Selmer groups
8.1 Notation and hypotheses
Letf, g be cuspidal modular newforms of weightsk+2, k+2, respectively, and levels
Nf, Ngprime top. Wedopermit here the casek= −1. We suppose, however, thatk>k,
so in particulark0, and we choose an integerjsuch thatk+1jk. Ifj= k+2k+1, then we assume thatεfεgis not trivial, whereεf andεgare the characters off andg.
As usual, we letE be a finite extension ofQpwith ring of integersO, containing the
coefficients off andg. Our goal will be to bound the Selmer group associated with the Galois representationMO(f⊗g)(1+j), in terms of theL-valueL(f, g,1+j); our hypotheses on (k, k, j) are precisely those required to ensure that thisL-value is acriticalvalue.
It will be convenient to impose the following local assumptions atp:
• (p-regularity) We haveαf =βf andαg =βg, whereαf,βf are the roots of the Hecke
polynomial off atp, and similarly forg. • (no local zero) None of the pairwise products
αfαg,αfβg,βfαg,βfβg
is equal topjorp1+j, so the Euler factor ofL(f, g, s) atpdoes not vanish ats = jor
s=1+j.
• (nobility offα) Iff is ordinary, then eitherαf is the unit root of the Hecke polynomial,
orME(f)|GQpis not the direct sum of two characters (so the eigenformfαis noble in
the sense of4.6.3).
• (nobility ofgα andgβ) Ifk 0, thenME(g)|GQp does not split as a direct sum of
characters, so bothp-stabilisationsgαandgβare noble.
Remark 8.1.1 (1) In our arguments we will use bothp-stabilisationsgαandgβofg, but only the onep-stabilisationfα off; in particular, we do not require that the other
p-stabilisationfβbe noble.
(2) Note that the “no local zero” hypothesis is automatic, for weight reasons, unlessk+k
is even andj= k+2k orj= k+2k +1 (so theL-valueL(f, g,1+j) is a “near-central” value).
Thep-regularity hypothesis implies that we have direct sum decompositions Dcris(ME(f)∗)=Dcris(ME(f)∗)αf ⊕Dcris(ME(f)∗)βf
where ϕ acts on the two direct summands as multiplication byαf−1,βf−1, respectively, and similarly forg. This induces a decomposition ofDcris(ME(f ⊗g)∗) into four direct summandsDcris(ME(f ⊗g)∗)αfαg etc.
Definition 8.1.2 We write