Let k be a field and A a commutative k-algebra. For every integer i⩾1 put Bi=A⊗kA⊗k⋯⊗kA and endow Bi with the structure of an A-algebra obtained by making A act on the first component. Put C1=B2,C2=B3,C3=B4⊕B3, and define A-linear mappings d2:C2→C1 and d3:C3→C2 by the formulas
¹ If A is a local, noetherian and complete algebra over a (discrete) field k, this definition coincides with that given in VIII, p. 98, exerc. 30, according to Exercise 5 below.
for every choice of elements x,y,z,α,β of A. Thus one obtains a complex of A-modules
Ck(A):C3d3C2d2C1.
If N is an A-module, we shall say that a k-bilinear mapping f:A×A→N is a 2-cocycle if one has the identity xf(y,z)−f(xy,z)+f(x,yz)−zf(x,y)=0 for x,y and z in A, and that it is symmetric if one has f(x,y)=f(y,x) for every pair (x,y)∈A×A.
a) The following conditions are equivalent: (i) the complex HomA(Ck(A),N) is acyclic for every A-module N; (ii) whatever the A-module N, every symmetric 2-cocycle f:A×A→N is a 1-coboundary, that is to say, of the form f(a,b)=ag(b)+bg(a)−g(ab) with g∈Homk(A,N); (iii) A is a smooth algebra over k.
b) If A is a field, the preceding conditions are also equivalent to the condition that the complex Ck(A) be acyclic.