The explicit long exact sequence in low degrees #
A short exact sequence 0 → A → B → C → 0 of discrete modules over a topological group G
induces the exact sequence
0 → H⁰(G, A) → H⁰(G, B) → H⁰(G, C) →δ⁰→ H¹(G, A) → H¹(G, B) → H¹(G, C) →δ¹→ H²(G, A) → H²(G, B)
of the explicit continuous-cochain model. This file proves exactness at each of its eight nodes.
The two leftmost nodes need no topology on G at all; the remaining six carry exactly the
hypotheses their maps require, so the three nodes touching H² also ask for a continuous
multiplication on G.
The arguments are the ordinary diagram chases, run on continuous cochains. The topological input
is entirely in ShortExact.lean: a continuous cochain into the discrete C lifts to a continuous
cochain into B, and a continuous cochain into B killed by the projection retracts to a
continuous cochain into A. Once a chase has produced a cochain, mem_Z1_of_incl_comp_mem_Z1
and mem_Z2_of_incl_comp_mem_Z2 are what put it back into the continuous cocycles.
Main statements #
TauCeti.ContCohomology.DiscreteShortExact.explicitLongExact_H0A: the coefficient inclusion is injective onH⁰.TauCeti.ContCohomology.DiscreteShortExact.explicitLongExact_H0B: exactness atH⁰(G, B).TauCeti.ContCohomology.DiscreteShortExact.explicitLongExact_H0C: exactness atH⁰(G, C).TauCeti.ContCohomology.DiscreteShortExact.explicitLongExact_H1A: exactness atH¹(G, A).TauCeti.ContCohomology.DiscreteShortExact.explicitLongExact_H1B: exactness atH¹(G, B).TauCeti.ContCohomology.DiscreteShortExact.explicitLongExact_H1C: exactness atH¹(G, C).TauCeti.ContCohomology.DiscreteShortExact.explicitLongExact_H2A: exactness atH²(G, A).TauCeti.ContCohomology.DiscreteShortExact.explicitLongExact_H2B: exactness atH²(G, B).TauCeti.ContCohomology.DiscreteShortExact.explicitDelta0_surjective_of_subsingletonandexplicitDelta1_bijective_of_subsingleton: whenH¹(G, B)andH²(G, B)vanish,δ⁰is onto andδ¹is bijective.TauCeti.ContCohomology.DiscreteShortExact.explicitCoeff0_surjective_of_subsingleton: whenH¹(G, A)vanishes,H⁰(G, B) → H⁰(G, C)is onto.
The maps and their normalization are those of
TauCeti/RepresentationTheory/Homological/ContCohomology/LowDegree.lean,
ExplicitFunctoriality.lean and ShortExact.lean. This implements the eight exactness nodes of
the long exact sequence milestone of Layer 5 of the human-authored roadmap at
TauCetiRoadmap/ProfiniteCohomology/README.md, whose Suggested.lean fixes the eight names
above.
References #
- J. Neukirch, A. Schmidt, K. Wingberg, Cohomology of Number Fields, 2nd ed., (1.3.2): the low-degree long exact sequence used here.
Coefficient maps on chosen representatives #
Every chase below hands a cocycle to a coefficient map and has to recognize the result as the
class of a cocycle already in hand. These two lemmas do that recognition once; they are the
explicitCoeff counterpart of explicitDelta0_apply and explicitDelta1_apply.
Exactness at H⁰(G, A): the coefficient inclusion remains injective on invariants.
Exactness at H⁰(G, B): the invariant elements killed by the projection are precisely
the invariant elements coming from A.
Exactness at H⁰(G, C): the image of the projection on invariants is the kernel of the
connecting homomorphism δ⁰ : H⁰(G, C) → H¹(G, A).
The degree-one nodes #
Exactness at H¹(G, A) and at H¹(G, B). Neither node mentions H², so neither needs a
continuous multiplication on G.
Exactness at H¹(G, A), the node where δ⁰ lands.
Exactness at H¹(G, B).
The nodes touching degree two #
Exactness at H¹(G, C), where δ¹ leaves, and at H²(G, A) and H²(G, B). All three mention
H², hence carry the continuous multiplication on G that makes d¹ preserve continuity.
Exactness at H¹(G, C), the node where δ¹ leaves.
Exactness at H²(G, A), the node where δ¹ lands.
Exactness at H²(G, B), the eighth and last node.
An acyclic middle term #
When B has vanishing H¹ and H², the long exact sequence makes the connecting maps shift
degree: δ⁰ is onto and δ¹ is an isomorphism. This is the step of the dimension-shifting
argument that does not depend on how the acyclic module was built.
If the middle term has vanishing H¹, then δ⁰ : H⁰(G, C) → H¹(G, A) is surjective.
If the middle term has vanishing H¹ and H², then δ¹ : H¹(G, C) → H²(G, A) is
bijective.
An acyclic first term #
When A has vanishing H¹, the connecting map δ⁰ is zero, so taking invariants is exact on the
right: every invariant of C is the image of an invariant of B.
If the first term has vanishing H¹, then H⁰(G, B) → H⁰(G, C) is surjective.