This thesis explores the potential decreolisation of Bislama, the creole of Vanuatu. It focuses on the evolution of the written language and consists in the study of a corpus of documents in Bislama in various genres and over a period of 40 years. The first three chapters of this thesis focus on the background, setting first the historical, geographical and sociological context of the study. In the second chapter, we present the theoretical frame on pidgins and creoles which shows that our study fills in a gap in the literature. Our work is diachronic in essence but also looks into the place of Bislama in the formal education system of Vanuatu. The methodology was therefore two-fold. The main part of the thesis is an empirical quantitative study of a database of written material that we constructed as well as other documents not included in the database. The part on education relies on interviews made in Vanuatu during my fieldtrips. The hypothesis of decreolisation is discussed in Chapters 4 and 5 and the present research shows that the anglicisation of written Bislama varies considerably between the genres. For some genres and in some of their sub-categories, there is evidence of decreolisation taking place. However, no generalisation can be made as in other genres the phenomenon is just not occurring in an obvious way or not at all. Chapter 6 summarises the conclusions on the results while it also highlights the probable link between the potential decreolisation of Bislama and the fact that it is neither taught nor used as a medium of instruction in Ni-Vanuatu schools.