MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top


in our lectures we have seen the following:

Let $(M, J)$ be a complex manifold, i.e., $J$ is an integrable almost complex structure. Then a real vector field $X$ is an infinitesimal automorphism of $J$ if and only if its component $X^{1,0}\in \Gamma(T^{1,0}M)$ is holomorphic.

Now let $(M, J)$ be an almost complex manifold and $X \in \chi(M)$ be an infinitesimal automorphism of $J$, i.e., $[X, JY]=J[X, Y]$ for all $Y \in \chi(M)$.

Can we then conclude that the mapping $X^{1,0}f$ with $X^{1,0}:=\frac{1}{2}(X-iJX)$ is pseudoholomorphic for every pseudoholomorphic mapping $f: M \to \mathbb{C}$?

Thank you and best regards,


share|cite|improve this question
Can someone edit the question a bit? It needs some clean-up and I don't have the power for that. – Spiro Karigiannis Nov 8 '10 at 21:53
Done. I hope it reads better now. – José Figueroa-O'Farrill Nov 8 '10 at 22:09
If $J$ is sufficiently non-integrable (meaning that the values of $N_J$ generate the tangent space), there are non non-constant $J$-holomorphic functions $f:M\to\mathbb{C}$, since $N_J(X,Y)f=0$ for all vector fields $X,Y$. – BS. Nov 9 '10 at 13:07
... but note that $N_J$ has at most a one (complex) dimensional image if $M$ has complex dimension 2, so maybe there are a few non-constant $J$-holomorphic functions in this case ("of one variable", so to speak). – BS. Nov 9 '10 at 13:12
up vote 6 down vote accepted

I can give you a partial answer- a reformulation of your question. Perhaps others will then be able to give a counterexample. (I think it is not true in general.) Let's start from the beginning.

A map $h : M \to \mathbb C$ is $J$-holomorphic iff $h_* (J Y) = i h_* (Y)$ for every vector field $Y$ on $M$. Equivalently, $dh(JY) = i dh(Y)$, or $(JY) (h) = i Y(h)$, or $(Y + i JY)h = 0$, or $(Y - iJY)h = 2Yh$ for any $Y$.

The condition that $X$ is an infinitesmal automorphism of the almost complex structure $J$ is that $[X,JY] = J[X,Y]$ for all $Y$. We want to investigate the question of under what conditions on $J$, for such an $X$, is it true that $(X- i JX)f = 2 Xf$ will be $J$-holomorphic for any $J$-holomorphic $f$.

Ignoring the factor of $2$, we see that \begin{equation*} Xf \text{ is $J$-holomorphic} \Leftrightarrow (Y + iJY)(Xf) = 0 \text{ for all $Y$} \end{equation*} \begin{equation*} \Leftrightarrow Y(Xf) = -i(JY)(Xf) \text{ for all $Y$} \end{equation*}

Recall the definition of the Nijenhuis tensor $N_J(X,Y) = [JX, JY] - J[JX, Y] - J[X,JY] - [X,Y]$. Since $X$ is an infinitesmal automorphism of $J$, the last two terms cancel. Applying both sides of this equation to a $J$-holomorphic map $f$, we get \begin{equation*} N_J(X,Y) f = (JX)((JY)f) - (JY)((JX)f) - (J[JX,Y])f = (JX)(iYf) - (JY)(iXf) - i [JX,Y]f \end{equation*} \begin{equation*} = i \left( (JX) (Yf) - (JY)(Xf) - (JX)(Yf) + Y((JX)f) \right) \end{equation*} \begin{equation*} = i \left( - (JY)(Xf) + i Y(Xf) \right) \end{equation*} which is exactly the quantity that we need to vanish.

Therefore, $Xf$ will also be $J$-holomorphic iff $N_J(X,Y)f = 0$ for all vector fields $Y$. In particular, this always holds if $J$ is integrable, since then $N_J$ vanishes identically. I don't know an example where $J$ is not integrable but where such an $X$ and $f$ exist. If there is one (which I would bet there is), then in that case $Xf$ would not be $J$-holomorphic.

share|cite|improve this answer
Okay, thanks... I assume that there is one, too... Now we only need someone you can construct an example :) – Differentialgeometer Nov 9 '10 at 11:49

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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