Eric Forgy
Noncommutative Stochastic Calculus

Given any first order calculus (d,Ω 1(A)) on a commutative algebra A, we have

d(fg)=d(gf)[df,g]=[dg,f].d(f g) = d(g f)\implies [d f,g] = [d g,f].

According to

this is enough to show that the commutator is a symmetric A-bilinear function of df and dg, i.e.

[df,g]=C(df,dg).[d f,g] = C(d f,d g).

The symmetric A-bilinear map C extends to a product

C:Ω 1(A)×Ω 1(A)Ω 1(A),C:\Omega^1(A)\times\Omega^1(A)\to\Omega^1(A),

since Ω 1(A) is generated by elements of the form a(db) for a,bA. We will denote this product

αβC(α,β).\alpha\bullet\beta \coloneqq C(\alpha,\beta).

This allows us to write the product rule in “left component form” as

d(fg)=f(dg)+g(df)+dfdg,d(f g) = f(d g) + g(d f) + d f\bullet d g,

in “right component form” as

d(fg)=(dg)f+(df)gdfdg,d(f g) = (d g)f + (d f)g - d f\bullet d g,

or the “symmetrized form” as

d(fg)=12[f(dg)+(dg)f]+12[g(df)+(df)g].d(f g) = \frac{1}{2}\left[f(d g) + (d g)f\right] + \frac{1}{2}\left[g(d f) + (d f)g\right].