5
$\begingroup$

I'm looking to prove that the pullback of a smooth function distributes over wedge product, i.e. $$\varphi^*(\omega \wedge \eta) = \varphi^* \omega \wedge \varphi^* \eta. $$

Here the question is answered. This is an excerpt from the answer:

$f^*(\omega\wedge \theta)(v_1,\cdots ,v_p,w_1,\cdots ,w_q)=(\omega\wedge \theta)(f_*(v_1),\cdots ,f_*(v_p),f_*(w_1),\cdots ,f_*(w_q)).$ using the summation formula for wedge product. Which will be same as $\omega(f∗(v_1),⋯,f∗(v_p))\wedge \theta(f∗(w_1),⋯,f∗(w_q)).$ Which is (by definition) same as $f_*(\omega)\wedge f_*(\theta).$

The part I have trouble with is the third line. $\omega(f∗(v_1),⋯,f∗(v_p))$ and $\theta(f∗(w_1),⋯,f∗(w_q))$ are real numbers, so how can we wedge them? There might be some abuse of notation here I'm not familiar with.

When I try to show $\varphi^*(\omega \wedge \eta) = \varphi^* \omega \wedge \varphi^* \eta $, my attempt breaks down when I try to simplify $(\varphi^*\omega \wedge \varphi^* \eta)(v_1, \dotsc, v_{p+q})$. Should I try to work with decomposable forms?

$\endgroup$
  • $\begingroup$ The wedge product of two $0$-forms (aka scalars or functions, depends on whether you look at a single point or not), is just the ordinary (pointwise) product. $\endgroup$ – Daniel Fischer Dec 15 '13 at 20:45
  • $\begingroup$ @DanielFischer A ha, and where do we get $$(\omega \wedge \theta)(f_*(v_1), \dotsc, f_*(v_{p+q}))=\omega(f_*(v_1), \dotsc, f_*(v_p))\wedge \theta(f_*(w_{1}), \dotsc, f_*(w_q))?$$ $\endgroup$ – Eric Auld Dec 15 '13 at 20:48
  • $\begingroup$ I don't think we do. $$(dx\wedge dy)\left(\frac{\partial}{\partial y},\frac{\partial}{\partial x}\right) \neq dx\left(\frac{\partial}{\partial y}\right) \cdot dy\left(\frac{\partial}{\partial x}\right).$$ We must permute and sum. $\endgroup$ – Daniel Fischer Dec 15 '13 at 20:53
  • $\begingroup$ OK, good. At least I'm not thinking about it wrong. How would you show $$(\omega \wedge \eta) (dF(v_1), \dotsc, dF(v_{k+l})) = (F^*\omega \wedge F^*\eta) (v_1, \dotsc, v_{k+l})?$$ Everywhere I'm seeing has the proof omitted, as if it's trivial. Should I perhaps try so show on decomposable forms $(\omega_1 \wedge \dotsb \wedge \omega_k) \wedge (\eta_1 \wedge \dotsb \wedge \eta_l)$? But that doesn't seems promising because I don't know how to take $\varphi^*((\omega_1 \wedge \dotsb \wedge \omega_k) \wedge (\eta_1 \wedge \dotsb \wedge \eta_l))$. $\endgroup$ – Eric Auld Dec 15 '13 at 20:59
  • $\begingroup$ @DanielFischer Oh, I think I'm seeing it. We have to permute and sum, but on each term, they are the same. $\endgroup$ – Eric Auld Dec 15 '13 at 21:08
1
$\begingroup$

The answer mentioned should have noted that we can say: \begin{equation} \begin{aligned} \varphi^*(\omega)(v_1, \dotsc, v_k)&\varphi^*(\eta)(v_{k+1}, \dotsc, v_{k+l}) \\&= \omega(dF^*(v_1), \dotsc, dF^*(v_k))\eta(dF^*(v_{k+1}), \dotsc, dF^*(v_{k+l})) \end{aligned} \end{equation} and we can extend this termwise result over the $\text{Alt}$ map.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.