Formal deformation theory in left-proper model categories (Q2280160)

The intention of this article is to make a foundation for work on deformation theory of algebraic schemes and their morphisms in the model category theory. The article starts out by explaining the underlying ideas: It is more or less an agreement that the deformation theory over a field of characteristic \(0\) is controlled by a differential graded Lie algebra. This includes the Maurer-Cartan modulus gauge action, and the agreement above is a theorem of Lurie-Pridham following from Lurie's axiomatic definition of a deformation problem in the language of higher category theory. The Lurie-Pridham theorem represents the deformation problem by a functor \(\Phi,\) and the homotopy class of the controlling DG-Lie algebra is reconstructed from the values of \(\Phi.\) In most cases, the main interest is in the controlling DG-Lie algbra in order to obtain information about \(\Phi.\) In this article, the main goal is to construct the controlling DG-Lie algebra starting from the underlying algebro-geometric data. This is achieved by studying explicit examples already existing, and attempting to unify them. For affine schemes, the controlling DG-Lie algebra is the DG-Lie algebra of derivations of a multiplicative Tate-Quillen resolution, where the Maurer-Cartan elements corresponds to perturbations of the differential. The same receipt extends to (non-positively graded) DG-affine schemes, giving a good notion of deformations of non-positively graded differential graded local artinian rings. The example from affine schemes indicates that for general separated schemes, the right DG-Lie algebra controlling deformations can be constructed by taking derivations of a Palamodov resolvent. This resolvent has a complicated combinatorial structure, and it is very hard to conclude with interesting results. In this article, the combinatorics of the Palamodov resolvent is interpreted as the property of being cofibrant in a suitable model category and using the various lifting and factorization axioms of model categories in order to provide proofs. This gives results and new information when the deformation theory of affine schemes is revisited in the framework of model categories, and the authors claim that the work following this fundamental article are going to generalize to not necessarily affine schemes. Every multiplicative Tate-Quillen resolution of a commutative algebra is a cofibrant replacement in the category \(\mathbf{CDGA}_{\mathbb K}^{\leq 0}\) of differential graded commutative algebras in non-positive degrees, equipped with the projective model structure. The notion of deformations is formulated in terms of the model structure, in particular the notion of flatness and thickening. The first result of the present paper a definition of a good formal deformation theory of a morphism on every model category in which every cofibration is flat. This property is satisfied for several left-proper model categories used in concrete applications, in particular \(\mathbf{CDGA}_{\mathbb K}^{\leq 0}\). The deformation theory of a morphism is homotopy invariant: this says that given morphisms \(K\overset f\rightarrow X\overset g\rightarrow Y\) with \(g\) a weak equivalence, \(f,g\) has equal deformation theories. Thus the attention can be restricted to deformations of cofibrations. The second main result tells that in the category \(\mathbf{CDGA}_{\mathbb K}^{\leq 0}\), the general notion of deformation is equivalent to the notion introduced by Hinich, and gives the classical notion of deformation when restricted to algebras concentrated in degree \(0\). The proof of this contains, and relies on, the fact that the lifting property and the (C-FW), (CW-F) factorization properties are unobstructed, meaning that they can be lifted along every surjective morphism of DG-local artin rings. The proof of the unobstructedness properties are quite involved, and is a consequence of a non-trivial technical result about liftings of trivial idempotents in cofibrant objects. The general theory is carried out on a fixed model category \(\mathbf{M},\) but the main relevant examples for the applications are the categories \(\mathbf{CDGA}_{\mathbb K}\) of differential graded commutative algebras over a field \(\mathbb K\) of characteristic \(0\) and the full subcategory \(\mathbf{CDGA}_{\mathbb K}^{\leq 0}\) of algebras concentrated in non-positive degrees, equipped with the projective model structure. The article is well written (with at lot of misprints though), and gives a very nice general framework for deformations of objects in model categories. The main general results are essential for later applications, and I look forward in excitement for the future work based on the general theory given in this article.