We consider operations between languages, based on splitting the underlying alphabet into two disjoint sets. Such operations are generalizations of the classical catenation or shuffle operation. With such operations rational and algebraic languages can be defined similar to the classical case. The basic properties of the corresponding language families are investigated.