Functoriality of the tuned ideal under smooth pull-back and change of fields #
[Kol07, Lemma 74 (4)]: "if f : Y → X is smooth, then D(f^* I) = f^*(D(I))"
(derivativeIter_comap_of_smooth). The maximal coefficient ideal
W_s(I) = ∑_{wt(e) ≥ s} ∏_j (D^j I)^{e_j} of [Kol07, Definition 98] is built from the derivative
ideals by sums, products and powers, all of which commute with pull-back (comap_iSup,
comap_finset_prod, comap_pow); hence W_s(g^* I) = g^* W_s(I) for a smooth morphism g : Y → X
of smooth k-schemes (W_comap_of_smooth), the structure morphism of Y being g ≫ f. This is
what Step 2.3 of [Kol07, 104] and Step 3 of the proof of [Kol07, Theorem 103] need; compare Wlo05,
Lemma 2.9.3 for the analogous statement about Włodarczyk's coefficient ideal.
The same computation, with the derivative ideals commuting with a change of the base field along
σ : k → L in characteristic zero (derivativeIter_comap_of_isPullback_specMap of
Hironaka.Scheme.IdealSheaf.Derivative.BaseChange), gives W_s(p^* I) = p^* W_s(I) on a cartesian
square p : X_L → X over Spec σ (W_comap_of_isPullback_specMap), the derivative ideals of X_L
being taken over L; this is the change-of-fields functoriality [Kol07, 34.2] for the tuned ideal.
The tuned ideal commutes with pull-back along a smooth morphism, W_s(g^* I) = g^* W_s(I)
([Kol07, Lemma 74 (4)] applied to the construction of [Kol07, Definition 98]).
The tuned ideal commutes with the change of fields along σ : k → L in characteristic zero: on
a cartesian square p : X_L → X over Spec σ, W_s(p^* I) = p^* W_s(I), the derivative ideals of
X_L taken over L ([Kol07, Lemma 74 (4)] and [Kol07, 34.2]; compare [Wlo05, Lemma 2.9.3]).