[guided]The point of this step is to convert holomorphicity of $f$ into a statement about the type decomposition of forms. For each $x\in X$, the differential of $f$ is the real-[linear map](/page/Linear%20Map)
\begin{align*}
df_x:T_xX\to T_{f(x)}Y.
\end{align*}
Because $f$ is holomorphic, this real-linear map is compatible with the two complex structures:
\begin{align*}
df_x(J_Xv)=J_Ydf_x(v)
\end{align*}
for every $v\in T_xX$.
Now fix $x\in X$ and tangent vectors $u,v\in T_xX$. The pullback form is defined by inserting the pushed-forward tangent vectors into $\omega$:
\begin{align*}
(f^*\omega)_x(u,v)=\omega_{f(x)}(df_x(u),df_x(v)).
\end{align*}
To check that $f^*\omega$ has type $(1,1)$ as a real two-form, we test invariance under applying $J_X$ to both arguments. Using the definition of pullback gives
\begin{align*}
(f^*\omega)_x(J_Xu,J_Xv)=\omega_{f(x)}(df_x(J_Xu),df_x(J_Xv)).
\end{align*}
Holomorphicity replaces each occurrence of $df_x(J_X\cdot)$ by $J_Ydf_x(\cdot)$, so
\begin{align*}
(f^*\omega)_x(J_Xu,J_Xv)=\omega_{f(x)}(J_Ydf_x(u),J_Ydf_x(v)).
\end{align*}
Since $\omega$ is a real $(1,1)$-form on $Y$, it satisfies
\begin{align*}
\omega_{f(x)}(J_Ya,J_Yb)=\omega_{f(x)}(a,b)
\end{align*}
for all $a,b\in T_{f(x)}Y$. Applying this with $a=df_x(u)$ and $b=df_x(v)$ gives
\begin{align*}
(f^*\omega)_x(J_Xu,J_Xv)=\omega_{f(x)}(df_x(u),df_x(v)).
\end{align*}
By the definition of pullback, the right-hand side is
\begin{align*}
(f^*\omega)_x(u,v).
\end{align*}
Therefore $f^*\omega$ has type $(1,1)$.
Reality is also preserved under pullback: since $\omega$ takes real values on real tangent vectors, the form $f^*\omega$ takes real values on real tangent vectors. Hence $f^*\omega$ is a real $(1,1)$-form on $X$. Together with the closedness already proved, the compact Kähler Hodge decomposition on $X$ [citetheorem:8066] identifies the de Rham class of this closed real $(1,1)$-form as an element of
\begin{align*}
H^{1,1}(X)\cap H^2(X,\mathbb R).
\end{align*}[/guided]