I am reading Guillemin and Pollack's Differential Topology. For the proof on Page 164, I was not able to get through the last step.

$$f^*(w \wedge \theta) = (f^*w) \wedge (f^* \theta)$$

Definition. If $f: X \to Y$ is a smooth map and $\omega$ is a $p$-form on $Y$, define a $p$-form $f^*\omega$ on $X$ as follows: $$f^*\omega(x) = (df_x)^*\omega[f(x)].$$

According to Daniel Robert-Nicoud's ansewr to $f: X \to Y$ is a smooth map and $\omega$ is a $p$-form on $Y$, what is $\omega[f(x)]$?

Locally, differential form can be written as $$\omega_\alpha(y) = \alpha(y)dx^{i_1}\wedge\ldots\wedge dx^{i_p}$$ with $\alpha$ a smooth function. Then $$f^*\omega_\alpha(x) = (df_x)^*[(\alpha\circ f)(x)dx^{i_1}\wedge\ldots\wedge dx^{i_p}].$$

We write $$\omega_\alpha(y) = \alpha(y)dx^{i_1}\wedge\ldots\wedge dx^{i_p},$$ $$\theta_\beta(y) = \beta(y)dx^{j_1}\wedge\ldots\wedge dx^{j_q}.$$

Hence, $$\omega \wedge \theta = (\alpha(y)dx^{i_1}\wedge\ldots\wedge dx^{i_p}) \wedge (\beta(y)dx^{j_1}\wedge\ldots\wedge dx^{j_q})$$

Following James S. Cook's very brilliant answer Pullback expanded form.

$$\omega \wedge \theta = \alpha(y) \beta(y) \sigma(I,J) dx^{k_1}\wedge\ldots\wedge dx^{k_{p+q}}$$

Rename $\gamma(y) = \alpha(y) \beta(y) \sigma(I,J)$, we get $$f^*(w \wedge \theta) = (df_x)^* [(\gamma \circ f)(x) dx^{k_1}\wedge\ldots\wedge dx^{k_{p+q}}].$$

But $$(f^*w) \wedge (f^* \theta) = ((df_x)^*[(\alpha\circ f)(x)dx^{i_1}\wedge\ldots\wedge dx^{i_p}])\wedge((df_x)^*[(\beta\circ f)(x)dx^{j_1}\wedge\ldots\wedge dx^{j_q}])$$

So here I got stuck - I don't really know how to move around $\alpha, \beta$ under $df^*$, to get close to the left hand side expression.

Thank you!~


I read your approach but I am giving another approach which might be helpful. Lets start with the definition. If $f:X\rightarrow Y$ is a map and $\omega$ is a $p$ form in $Y$ then its pull back is defined as $f^*(\omega)v=\omega (f_*v).,$ where $f_*v=v(f),$ for any $v\in T_pX.$ Therefore Expand (assuming $\omega$ is a p form and $\theta$ is a q form) the term (where $v_i$and $w_i$ are in $T_pX$ ) $f^*(\omega\wedge \theta)(v_1,\cdots ,v_p,w_1,\cdots ,w_q)=(\omega\wedge \theta)(f_*(v_1),\cdots ,f_*(v_p),f_*(w_1),\cdots ,f_*(w_q)).$ 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 $f^*(\omega)\wedge f^*(\theta).$

  • $\begingroup$ looks great, thanks tessellation. I wonder if my approach is wrong. $\endgroup$ – WishingFish Jul 31 '13 at 6:01
  • 1
    $\begingroup$ @WishingFish notice he is doing what I said in the other thread. That is what the $df_x^*$ intends. It might be good to look back at math.stackexchange.com/questions/455233/… or math.stackexchange.com/questions/451246/pullback-of-a-1-form/… where I focus more on the issue of how the pull-back replaces the coefficient with the composition of the old coefficient. $\endgroup$ – James S. Cook Jul 31 '13 at 6:08
  • $\begingroup$ It feels like physicists and geometers haven't met an agreement on many mathematical expressions yet, but I am really confused and frustrated by the different notation from the one I am introduced. Because I am constrained in the idea that if I want to use another definition, I shall prove its equivalence first... @JamesS.Cook $\endgroup$ – WishingFish Jul 31 '13 at 16:38
  • $\begingroup$ To be more specific, I am confused with tessellation's definition $f^*(\omega)v = \omega(f_*v)$ with my $f^*(\omega)v = (df_x)^*\omega[f(x)]$, @JamesS.Cook $\endgroup$ – WishingFish Jul 31 '13 at 16:59
  • $\begingroup$ This answer may be misleading...see here math.stackexchange.com/questions/608269/… $\endgroup$ – Eric Auld Dec 15 '13 at 20:54

By definition, $$f^*(w \wedge \theta) = (df_x)^*(\omega\wedge\theta)[f(x)].$$

Page 162: If $\omega$ is a $p$-form and $\theta$ is a $q$-form, then $p+q$ form $$(\omega \wedge \theta) (x) = \omega(x) \wedge \theta(x).$$

Therefore, $$(df_x)^*(\omega\wedge\theta)[f(x)] = (df_x)^*\omega[f(x)] \wedge \theta[f(x)].$$

Page 159: $A^*$ is linear and that $$A^*(T \wedge S) = A^*T \wedge A^*S.$$

Therefore, $$(df_x)^*\omega[f(x)] \wedge \theta[f(x)] = (df_x)^*\omega[f(x)] \wedge (df_x)^*\theta[f(x)].$$

That is exactly $$(f^*w) \wedge (f^* \theta).$$


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.