Concerning Medieval French, text-processing has to handle two major problems: first the under-developed state of linguistic knowledge of scholars; secondly the very large variations in forms and spellings occurring throughout the manuscripts. In order to extract and provide lexicographical and grammatical information from many yet undescribed texts, our research group has decided to devise an expert system which would help identify and parse the text-words and would be linked to an open relational database. The design of our expert system is outlined here: we especially outline the initial knowledge base and the nature of the different sets of rules we will make use of. Some examples describe how the system will operate and the types of advantages that are expected from its use. Finally the paper lists the studies and published work which in the last years have provided the groundwork for this project.