Cotangent cohomology of a category of deformations (Q1817234)

There are two main basic problems in deformation theory of algebraic structures, namely the equivalence and integrality problems. The equivalence problem is the determination of whether a deformation is equivalent (by a `twisting') to the original structure, while the integrality problem deals with the extendability of a deformation defined up to \(n\)-th order in the deformation parameter, to higher orders. Much previous work in deformation theory has involved the construction, individually for different algebraic structures, of appropriate cohomology theories in which the obstruction to deformation may be found. For example, for associative algebras, the third Hochschild homology contains the deformation obstruction, while the bialgebra cohomology of Gerstenhaber and Schack is where the obstruction lies for bialgebras. The main result of this paper is a general construction of a cohomology theory for algebraic objects in any linear equationally given theory. That is, one is considering all algebraic structures defined by a vector space, \(V\), along with certain operations \(m_i: V^{\otimes s} \to V^{\otimes t}\) satisfying certain axioms; a theory is the structure containing the types of operations and the axioms. Those theories involving only structures for maps with \(t=1\) are known as algebraic and the deformation theory problem is solved for them using the triple cohomology [see \textit{T. F. Fox}, J. Pure Appl. Algebra 84, No. 1, 17-41 (1993; Zbl 0772.18006)]. After introducing general notions of a theory and core, along with the free theory \({\mathcal F} (X)\) generated by a core \(X\) (a theory without relations), quotients of theories by relations, modules, the free module \(A\langle X\rangle\) on a core \(X\), and the space of derivations, the paper defines a small complex, in whose cohomology lies an obstruction to deformations. This construction is based on a resolution \[ A\leftarrow {\mathcal F} (X) \leftarrow {\mathcal F} (X)\langle Y\rangle \leftarrow {\mathcal F} (X)\langle Z \rangle \] of the theory \(A \), and the small complex generated is \[ 0\to\text{Der} \bigl({\mathcal F} (X),M \bigr)\to \Hom_{{\mathcal F} (X)} \bigl({\mathcal F} (X)\langle Y\rangle, M\bigr) \to\Hom_{{\mathcal F} (X)} \bigl({\mathcal F} (X) \langle Z \rangle, M \bigr). \] For the examples of differential spaces, associative algebras and bialgebras, the obstruction is seen to reduce to those previously known. Unfortunately it is usually rather hard to prove that one has a resolution for a theory, but one important property of the author's construction is that it is defined from any pre-resolution, and that a systematic procedure exists for generating pre-resolutions which give nontrivial (meaning not too large) obstruction spaces. Indeed this procedure generates the same obstruction spaces as by the author and \textit{J. D. Stasheff} [J. Algebra 170, No. 1, 122-155 (1994; Zbl 0820.17020)]. For the case of Drinfel'd algebras (quasi-Hopf algebras) it is shown that the deformation theory so generated is identical with that developed by the author and \textit{S. Shnider} [Int. Math. Res. Not. 1994, No. 4, 169-176 (1994; Zbl 0821.16043); Trans. Am. Math. Soc. 348, No. 9, 3505-3547 (1996)].