Harmonic algebraic curves and noncrossing partitions (Q878070)

The authors recall that Gauss's first proof of the Fundamental Theorem of Algebra consists in proving that the harmonic real algebraic curves \[ R=\{z\in \mathbb C: \mathfrak{Re} f(z)=0\} \quad \text{and}\quad I=\{z\in \mathbb C: \mathfrak{Im} f(z)=0\} \] associated to a degree \(n\) polynomial \(f\in \mathbb C[T]\) have nonempty intersection. Indeed, for a large enough radius, each of the curves \(R\) and \(I\) meets the boundary \(S_r\) of the disk \(D_r\) of radius \(r\) and centered at the origin in exactly \(2n\) points, in such a way that the points of \(R\cap S_r\) alternate with those of \(I\cap S_r\) around \(S_r\). Gauss argued that \(R\cap D_r\) and \(I\cap D_r\) consist of \(n\) components and that each component of \(R\) must cross some component of \(I\) inside \(D_r\); hence \(R\cap I\) has at least \(n\) points, and so it has exactly \(n\) points. The article under review is a very careful analysis of the topology of the curves \(R\) and \(I\), both singly and together. The intersection of \(R\cup I\) with a suitable circle \(S_r\) looks like (the black lines on the surface of) a basketball, a notion which is precisely defined in the paper, and the main result is to prove the Inverse Basketball Theorem (Theorem 3.1) that says that each basketball arises from a polynomial \(f\in \mathbb C[T]\). A more subtle combinatorial-topological invariant of the polynomial \(f\) arises from the curves \[ C_{\theta}(f)=\{z\in\mathbb C: \mathfrak{Im}(e^{-i\theta}f(z)=0\} \] regarded as a family parametrized by \(\theta \in \mathbb R/\pi \mathbb Z\). Note that \(I=C_0(f)\) and \(R=C_{\pi/2}(f)\). The authors gives some first steps in the study of this object. With the words of the autors, there appears to be much more to say from both the combinatorial and geometric points of view.