Prime and composite Laurent polynomials (Q1041171)

This paper studies rational functions from the Riemann sphere to itself, and in particular their representation as compositions of lower-degree rational functions: for a rational function \(f\), a decomposition \(f=F_r\circ F_{r-1}\circ \dots \circ F_1\) is called ``maximal'' if all \(F_i\) have degree at least \(2\) and are not further decomposable. Such a decomposition of \(f\) need not be unique, even up to natural equivalence, and the goal of the paper is a classification of such decompositions. Such a theory exists for the case that \(f\) is a polynomial (due to \textit{J. Ritt} [American M. S. Trans. 23, 51--66 (1922; JFM 48.0079.01)]). The paper under review extends this theory to certain classes of rational functions, including those with at most two poles.