0

I have a basic differential form problem that I have not understood for many years. Resources of exterior derivative refer to Wikipedia Exterior Derivative.

My problem is

(1) Why it is said that the vector basis is $\displaystyle e_j=\frac{\partial}{\partial x^j}\, $ ?

(2) Why it is said that the dual vector basis is $e^j=dx^j\,$?

Thank you a lot for your help. I would be appreciated if you can prove them to me.

Kurt G.
  • 14,198
  • Why? Do you understand what a vector space and its dual space are, and what a basis and a dual basis are? If you do, then you know the answers to your question (without the manifolds). If not, why are we talking about this? – Ted Shifrin Oct 10 '22 at 01:02
  • Maybe relevant: https://math.stackexchange.com/questions/3898940/tangent-vectors-as-directional-derivatives-on-manifolds – Miguel Oct 15 '22 at 10:09

1 Answers1

2

I think the statements (1) and (2) can only be motivated but not proved. The introduction of differential forms by Elie Cartan was in my opinion one of the most ingenious and fruitful ideas in mathematics of the twentieth century. The heart of the idea must be the well known chain rule (here in two dimensions): $$\tag{A} \frac{df(x(t),y(t))}{dt}=\frac{\partial f}{\partial x}\frac{dx}{dt}+\frac{\partial f}{\partial y}\frac{dy}{dt}\,. $$ Recall that a vector in $\mathbb R^2$ can be written as $$\tag{B} \boldsymbol{v}=v_1\boldsymbol{e}_1+v_2\boldsymbol{e}_2 $$ where $\boldsymbol{e}_i$ are the basis vectors, and $(v_1,v_2)$ the components of $\boldsymbol{v}\,.$ Since $t\mapsto (x(t),y(t))$ is a curve whose tangent vector has components $(\frac{dx}{dt},\frac{dy}{dt})$ we can write the RHS of (A) as $$ \boldsymbol{v}(f) $$ when we view $\frac{\partial}{\partial x}$ and $\frac{\partial}{\partial y}$ as the basis vectors $\boldsymbol{e}_1$ and $\boldsymbol{e}_2\,$.

If we "drop" $dt$ in (A) we obtain $$ df=\frac{\partial f}{\partial x}\,dx+\frac{\partial f}{\partial y}\,dy\,. $$ which can be seen as a dual relationship between $(\frac{\partial}{\partial x},\frac{\partial}{\partial y})$ and $(dx,dy)$. More precisely, $$\tag{C} dx(\textstyle\frac{\partial}{\partial x})=1\,,\quad dy(\textstyle\frac{\partial}{\partial y})=1\,,\quad dx(\textstyle\frac{\partial}{\partial y})=0\,,\quad dy(\textstyle\frac{\partial}{\partial x})=0\,. $$ With $\boldsymbol{v}=v_1\frac{\partial}{\partial x}+v_2\frac{\partial}{\partial y}$ this leads directly to $$ df(\boldsymbol{v})=\frac{\partial f}{\partial x}v_1+\frac{\partial f}{\partial y}v_2=\boldsymbol{v}\cdot\nabla f $$ which is the directional derivative of $f$ in the direction of $\boldsymbol{v}\,.$ The dual relationships (C) are now seen as nothing else than describing the simple fact that

  • the directional derivative of $f(x,y)=x$ in the direction of $\frac{\partial}{\partial x}$ is one and zero in the direction of $\frac{\partial}{\partial y}\,,$ and so on.

I stop here because there are much better pedagogical introductions to differential forms. My favourite one is pp. 53 in the book Gravitation by Misner, Thorne and Wheeler.

Kurt G.
  • 14,198
  • Since =<, ∂/∂> , f is a zero form, df is a one form, but <, ∂/∂> is a real number, why <, ∂/∂> is a one form by the way? Here <, ∂/∂> is the dot product for x is a vector. – inv inv Oct 10 '22 at 19:03
  • 1
    No dot product. Evaluating an element of $V^$ on an element of $V$ gives a scalar (or evaluating* a $1$-form on a vector field gives a function). – Ted Shifrin Oct 10 '22 at 19:38
  • Not sure what you meant... – inv inv Oct 12 '22 at 04:11