Is ths set of vector fields on a manifold a ring

differential-geometryriemannian-geometryVector Fields

I´m studying differential geometry and we know that if $T$ is the set of smooth vector fields defined on a manifold $M$, then we can see that $T$ is a $R-$module with $R$ the set of smooth real-valued functions defined on $M$.

However, Can we give to $T$ a ring structure with the "multiplication" defined as $VW(f)=V(W(f))$ por all $f \in R$? Is it true that, if $V$ and $W$ are vector fields defined on a neighborhood $U$, then $VW$ is defined on $U$? I cannot see any counterexample for this.

Is true in general or not?

Greetings!

Best Answer

The composition $VW$ will be, in general, a non-linear differential operator, not a vector field.

For example, $\frac{\partial^2}{\partial x^2}$ is not a "derivation" in the sense that it doesn't verify the product rule of derivatives.

So, $VW$ is defined, of course, but it's not a vector field. This "product" is not an operation: it doesn't produce another vector field. It is a product in the algebra of differential operators, though.

On the other hand, the commutator of this product: $[V, W] = VW - WV$, called the Lie bracket of vector fields, is a linear differential operator: a vector field. This turns the module of vector fields into a type of noncommutative ring called Lie algebra. This is very important in the geometrical study of differential equations (the flow associated to $[V,W]$ is related with the flows of $V, W$ in an interesting way).

Related Question