The definition I have seen for an almost complex structure is the following

$$J:TM\to TM$$ which is linear fibre by fibre, such that $J^2 = -\text{Id}$, and such that $\pi(J(X_x)) = x$ where $\pi$ is the projection of $TM$ on $M$.

I see that this implies a map $$\tilde{J}:\mathfrak{X}(M)\to\mathfrak{X}(M)$$ such that $(\tilde{J})^2 = -\text{Id}$ and given by $\tilde{J}(X)(x) = J(X_x)$

My question is:

Given a linear map $\tilde{J}:\mathfrak{X}(M)\to\mathfrak{X}(M)$ such that $(\tilde{J})^2 = -\text{Id}$, can we reconstruct an almost complex structure on $M$?

  • $\begingroup$ What is the definition of $J(X_x)$ in $J(X)(x) = J(X_x)$? $\endgroup$ – user99914 Jun 8 '18 at 10:30
  • $\begingroup$ Okay, I'll just. it isn't very well written $\endgroup$ – tomak Jun 8 '18 at 10:32
  • 2
    $\begingroup$ Are there more properties assumed? For now $\tilde J$ might not preserve the fiber, and might not be linear on the fiber. $\endgroup$ – user99914 Jun 8 '18 at 10:36
  • $\begingroup$ The problem is that I don't know. I can go from $J$ to $\tilde{J}$, but I don't know the properties that I would need to be able to go the other way. So if there are properties that could be added to $\tilde{J}$ (that $\tilde{J}$ does indeed have if it comes from some $J$) we can add them. $\endgroup$ – tomak Jun 8 '18 at 11:55
  • $\begingroup$ for instance, it would be a good idea to add linearity to $\tilde{J}$ $\endgroup$ – tomak Jun 8 '18 at 12:00

Let $V$ be an infinite-dimensional vector space. By the axiom of choice, there is a basis $\{v_i\}_{i\in I}$ for $V$. As $I$ is infinite, we can find $A, B \subset I$ with $A\cap B = \emptyset$, $A\cup B = I$, and $|A| = |B|$. Let $\sigma : A \to B$ be a bijection. Then we can define a linear map $\tilde{J} : V \to V$ by setting $\tilde{J}(v_a) = v_{\sigma(a)}$ and $\tilde{J}(v_b) = -v_{\sigma^{-1}(b)}$; note that $\tilde{J}^2 = -\operatorname{id}_V$.

Now note that $\mathfrak{X}(M)$ is an infinite-dimensional vector space if $\dim M > 0$. If we could construct a $J$ from $\tilde{J}$, we would see that every positive-dimensional manifold admits an almost complex structure which is absurd.


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.