No label defined (Q3972289)

[For the entire collection see Zbl 0706.00019.] A brief account of this work has appeared in an earlier article [the author and \textit{N. Ercolani}, Phys. Lett. A 119, 112-116 (1986)]and an expanded version will follow. These lectures are merely intended to illustrate, in the simplest possible terms, our approach to the Painlevé problem. Precise statements of theorems, in the few instances where we have them, as well as nontrivial examples are all reserved for the work of \textit{N. Ercolani} and the author [What is integrability, Springer Ser. Nonlinear Dyn., 63-72 (1991; Zbl 0733.34008)]. We will consider only polynomial systems of equations, \((\cdot\equiv d/dt)\), (1) \(\dot x_ i=f_ i(\{x_ j\})\), \(i,j=1,\dots,n\), so that the right hand side does not introduce any singularities into the time flows. In fact, the differential equations define the continuation of \(x_ i(t)\) from real to complex times. All variables will henceforth be understood as complex. The Painlevé conjecture then asserts that if the singularities of \(x_ i(t)\) are no worse than poles for all \(t\) (i.e., no branch points, logarithms, essential singularities etc.) then (1) is ``integrable''. In other terms, a system whose solutions are globally meromorphic is said to possess the Painlevé property.