Decomposition of \(p\)-adic meromorphic functions (Q1898741)

Let \(f, f_1, f_2, \dots, f_n\) be meromorphic functions such that \(f= f_1 \circ f_2\circ \dots \circ f_n\). This is called a decomposition (or factorization) of \(f\), and \(f\) is called indecomposable (or prime) if no non-trivial decomposition exists. One may ask when such decompositions exist if and if so, whether they are unique, etc. For polynomials a fairly complete answer has been given by \textit{J. F. Ritt} [Trans. Am. Math. Soc. 23, 51-66 (1922; JFM 48.0079.01)]. An analogue for rational functions is not known. Concerning the decomposition of transcendental entire and meromorphic functions, there is a large amount of literature, but a systematic theory has not been developed yet. In this paper the authors address the above questions for \(p\)-adic meromorphic functions. Let \(A(C_p)\) be the ring of entire functions in \(C_p\) and \(M(C_p)\) be a field of meromorphic functions in \(C_p\). We only quote some typical results: Theorem 1.10 says that if \(F\in A(C_p) \setminus C_p [x]\) and \(\lambda\) is a positive integer such that on infinitely many circles of center 0 in \(C_p\) the function \(F\) has been 1 and \(\lambda\) zeros, then in any decomposition \(F= H\circ G\) with \(H,G\in A(C_p)\) one of the functions \(H\) and \(G\) is a polynomial of degree at most \(\lambda\). This leads to criteria for functions to be indecomposable. Theorem 2.2 says that if \(F= H\circ G\) with \(F\in A(C_p)\) and \(H, G\in M(C_p)\), then there exists an equivalent decomposition \(F= h\circ g\) with \(h, g\in A(C_p)\). It is known, and mentioned by the authors, that this latter result is false in the complex case. Finally, the authors discuss in which cases a function and some derivative of it can have factorizations with a common right factor.