In this article I show three methods to use the Negerhollands computer corpus for a better understanding of the eighteenth century variety of Negerhol lands. The first one is the study of metalinguistic comments which are taken up in several manuscripts. The second method is stemma/variety research between and within texts with the same content. The third method to study the do's and doubts of the first translators in order to present the right variety of the language is the analysis of annotations, changes and erasures of the writers. This philolog ical approach was common in Medieval studies, for instance, but is hardly used with respect to Creole languages.