Conditional i.i.d.-ness of a contractable process, via L² #
A contractable process on a standard Borel state space is conditionally i.i.d. with
directingProbabilityMeasure μ X as its directing measure:
ConditionallyIIDWith μ X (directingProbabilityMeasure μ X).
This is where the L² route meets the common ending. ViaL2/BlockFactorization.lean supplies the
conditional factorization of an indicator block given tailProcess X, and
conditionallyIIDWith_of_measure_inter_blockCylinder_eq_setLIntegral asks for that identity
integrated over a directing-measure event. The integration itself is not done here: it is
measure_inter_blockCylinder_eq_setLIntegral_of_condExp, which takes the factorization as a
hypothesis and is shared with the martingale route. This file only supplies the L² factorization
to it, for an arbitrary strictly monotone selection.
The witness is preserved: the conclusion names directingProbabilityMeasure μ X itself rather
than asserting an unnamed existential.
Like the factorization it consumes, this file reaches conditional i.i.d.-ness without a reverse
martingale: neither DeFinetti.BlockFactorization nor TailFactorization, JointRectangle,
DeFinetti.Theorem or anything under Probability.Martingale is in its transitive imports. The
shared integration lemma is neutral for the same reason — it proves no factorization, only
consumes one.
References #
- Roadmap:
TauCetiRoadmap/Exchangeability/README.md, Layer 3 — the martingale-free standard-Borel de Finetti route,deFinetti_viaL2. - The integration step is shared, not route-specific:
measure_inter_blockCylinder_eq_setLIntegral_of_condExpinDeFinetti/DirectingMeasure/BlockCylinder.leanserves both this route and the prefix selection inDeFinetti/JointRectangle.lean.
A contractable process on a standard Borel space is conditionally i.i.d., with
directingProbabilityMeasure μ X as its directing measure.
This conditions on the directing measure itself. It does not assert conditional independence given
the whole of tailProcess X, which would additionally require identifying the σ-algebra generated
by directingProbabilityMeasure μ X with the tail.