Lemma 98.3.1. The functor p : \mathcal{F} \to \mathcal{C}_\Lambda defined above is a predeformation category.
98.3 Predeformation categories
Let S be a locally Noetherian base scheme. Let
be a category fibred in groupoids. Let k be a field and let \mathop{\mathrm{Spec}}(k) \to S be a morphism of finite type (see Morphisms, Lemma 29.16.1). We will sometimes simply say that k is a field of finite type over S. Let x_0 be an object of \mathcal{X} lying over \mathop{\mathrm{Spec}}(k). Given S, \mathcal{X}, k, and x_0 we will construct a predeformation category, as defined in Formal Deformation Theory, Definition 90.6.2. The construction will resemble the construction of Formal Deformation Theory, Remark 90.6.4.
First, by Morphisms, Lemma 29.16.1 we may pick an affine open \mathop{\mathrm{Spec}}(\Lambda ) \subset S such that \mathop{\mathrm{Spec}}(k) \to S factors through \mathop{\mathrm{Spec}}(\Lambda ) and the associated ring map \Lambda \to k is finite. This provides us with the category \mathcal{C}_\Lambda , see Formal Deformation Theory, Definition 90.3.1. The category \mathcal{C}_\Lambda , up to canonical equivalence, does not depend on the choice of the affine open \mathop{\mathrm{Spec}}(\Lambda ) of S. Namely, \mathcal{C}_\Lambda is equivalent to the opposite of the category of factorizations
of the structure morphism such that A is an Artinian local ring and such that \mathop{\mathrm{Spec}}(k) \to \mathop{\mathrm{Spec}}(A) corresponds to a ring map A \to k which identifies k with the residue field of A.
We let \mathcal{F} = \mathcal{F}_{\mathcal{X}, k, x_0} be the category whose
objects are morphisms x_0 \to x of \mathcal{X} where p(x) = \mathop{\mathrm{Spec}}(A) with A an Artinian local ring and p(x_0) \to p(x) \to S a factorization as in (98.3.0.1), and
morphisms (x_0 \to x) \to (x_0 \to x') are commutative diagrams
\xymatrix{ x & & x' \ar[ll] \\ & x_0 \ar[lu] \ar[ru] }in \mathcal{X}. (Note the reversal of arrows.)
If x_0 \to x is an object of \mathcal{F} then writing p(x) = \mathop{\mathrm{Spec}}(A) we obtain an object A of \mathcal{C}_\Lambda . We often say that x_0 \to x or x lies over A. A morphism of \mathcal{F} between objects x_0 \to x lying over A and x_0 \to x' lying over A' corresponds to a morphism x' \to x of \mathcal{X}, hence a morphism p(x' \to x) : \mathop{\mathrm{Spec}}(A') \to \mathop{\mathrm{Spec}}(A) which in turn corresponds to a ring map A \to A'. As \mathcal{X} is a category over the category of schemes over S we see that A \to A' is \Lambda -algebra homomorphism. Thus we obtain a functor
We will use the notation \mathcal{F}(A) to denote the fibre category over an object A of \mathcal{C}_\Lambda . An object of \mathcal{F}(A) is simply a morphism x_0 \to x of \mathcal{X} such that x lies over \mathop{\mathrm{Spec}}(A) and x_0 \to x lies over \mathop{\mathrm{Spec}}(k) \to \mathop{\mathrm{Spec}}(A).
Proof. We have to show that \mathcal{F} is (a) cofibred in groupoids over \mathcal{C}_\Lambda and (b) that \mathcal{F}(k) is a category equivalent to a category with a single object and a single morphism.
Proof of (a). The fibre categories of \mathcal{F} over \mathcal{C}_\Lambda are groupoids as the fibre categories of \mathcal{X} are groupoids. Let A \to A' be a morphism of \mathcal{C}_\Lambda and let x_0 \to x be an object of \mathcal{F}(A). Because \mathcal{X} is fibred in groupoids, we can find a morphism x' \to x lying over \mathop{\mathrm{Spec}}(A') \to \mathop{\mathrm{Spec}}(A). Since the composition A \to A' \to k is equal the given map A \to k we see (by uniqueness of pullbacks up to isomorphism) that the pullback via \mathop{\mathrm{Spec}}(k) \to \mathop{\mathrm{Spec}}(A') of x' is x_0, i.e., that there exists a morphism x_0 \to x' lying over \mathop{\mathrm{Spec}}(k) \to \mathop{\mathrm{Spec}}(A') compatible with x_0 \to x and x' \to x. This proves that \mathcal{F} has pushforwards. We conclude by (the dual of) Categories, Lemma 4.35.2.
Proof of (b). If A = k, then \mathop{\mathrm{Spec}}(k) = \mathop{\mathrm{Spec}}(A) and since \mathcal{X} is fibred in groupoids over (\mathit{Sch}/S)_{fppf} we see that given any object x_0 \to x in \mathcal{F}(k) the morphism x_0 \to x is an isomorphism. Hence every object of \mathcal{F}(k) is isomorphic to x_0 \to x_0. Clearly the only self morphism of x_0 \to x_0 in \mathcal{F} is the identity. \square
Let S be a locally Noetherian base scheme. Let F : \mathcal{X} \to \mathcal{Y} be a 1-morphism between categories fibred in groupoids over (\mathit{Sch}/S)_{fppf}. Let k is a field of finite type over S. Let x_0 be an object of \mathcal{X} lying over \mathop{\mathrm{Spec}}(k). Set y_0 = F(x_0) which is an object of \mathcal{Y} lying over \mathop{\mathrm{Spec}}(k). Then F induces a functor
of categories cofibred over \mathcal{C}_\Lambda . Namely, to the object x_0 \to x of \mathcal{F}_{\mathcal{X}, k, x_0}(A) we associate the object F(x_0) \to F(x) of \mathcal{F}_{\mathcal{Y}, k, y_0}(A).
Lemma 98.3.2. Let S be a locally Noetherian scheme. Let F : \mathcal{X} \to \mathcal{Y} be a 1-morphism of categories fibred in groupoids over (\mathit{Sch}/S)_{fppf}. Assume either
F is formally smooth on objects (Criteria for Representability, Section 97.6),
F is representable by algebraic spaces and formally smooth, or
F is representable by algebraic spaces and smooth.
Then for every finite type field k over S and object x_0 of \mathcal{X} over k the functor (98.3.1.1) is smooth in the sense of Formal Deformation Theory, Definition 90.8.1.
Proof. Case (1) is a matter of unwinding the definitions. Assumption (2) implies (1) by Criteria for Representability, Lemma 97.6.3. Assumption (3) implies (2) by More on Morphisms of Spaces, Lemma 76.19.6 and the principle of Algebraic Stacks, Lemma 94.10.9. \square
Lemma 98.3.3. Let S be a locally Noetherian scheme. Let
be a 2-fibre product of categories fibred in groupoids over (\mathit{Sch}/S)_{fppf}. Let k be a finite type field over S and w_0 an object of \mathcal{W} over k. Let x_0, z_0, y_0 be the images of w_0 under the morphisms in the diagram. Then
is a fibre product of predeformation categories.
Proof. This is a matter of unwinding the definitions. Details omitted. \square
Comments (0)