Presentation /en

The Reflex project aims to provide the scientific community with a Reference Lexicon for African languages, as well as various tools for processing and analyzing this corpus. A more detailed description is available in
PDF format.

 This project follows on from the Sumale project, which aimed to bring together a vast collection of lexical data on African languages. Sumale was the LLACAN equivalent of the Sahélia project developed by Robert Nicolaï in Nice. Both use the Mariama software, developped by Robert Nicolaï and Jeanne Zerner.

Today, the goals are both more realistic and more ambitious. More realistic, because there is no question of achieving perfectly labelled lexicons organised hierarchically; more ambitious, because the aim now is to make as much data as possible available to the entire scientific community, by means of a simple and intuitive web interface.


One benefit of this project is its immediate feasibility. Many keyboarded lexicons already exist, they just need to be published online. The consultation interface is also readily available. It is in the spirit of the project to be able to upload incomplete or fragmentary data, pending later review.

Eventually, RefLex will be a reference corpus for African languages, freely available and usable by researchers. This goal assumes that the original data sources are fully respected. The principle is that all researchers can view the same document, old or not, under the same conditions.


Initially, we will post two types of documents:

  • Old documents that are in the public domain, such as Wintz’s Joola dictionary, published in 1909, or Clarke’s Limba dictionary, published in 1922.
  • Recent, first-hand documents, published or otherwise, whose distribution has been explicitly authorized by the author. This category includes for example Segerer’s two Bijogo lexicons (only one has been published) and Nougayrol’s Day lexicon.

In the second phase, we will strive to obtain copyright permission for any documents that are not in one or other of the above categories. The corresponding data will be published online.