Rut Foundation has launched a research project in collaboration with the Italian Geographical Society for the construction of a bilingual Somali-Italian computational lexicon dedicated to the era of Italy’s colonialist expansion in Africa.

The first phase of the project is focusing on the extraction and systematization of terms taken from the notebooks written by the explorer Ugo Ferrandi. The notebooks constitute a veritable mine of linguistic information: to describe objects, concepts and phenomena typical of Somali culture, Ferrandi reports a plethora of indigenous terms, thus portraying an ancient phase of the Somali language as it was spoken by shepherds and nomadic farmers in the period pre-colonial, before the advent of the European powers.

The first phase of the project is focusing on the extraction and systematization of the terms reported by Ferrandi in his notebooks. When possible, the description of words is enriched with information taken from the Somali lexicon incorporated into the Somali Corpus created by Jama Musse Jama (2006). The annotated and balanced corpus (includes texts of both poetry and literary and scientific prose) incorporates, in fact, a Somali lexicon with linguistic information relating to the words present in the corpus, such as frequency, collocations, etymology, synonyms and antonyms, spelling variants, definitions taken from a list of reference dictionaries as well as translations into English, Italian, French and Swedish.

However, since the corpus lexicon is encoded in a proprietary format, the connection with our termino-ontological resource cannot be made without the previous conversion of the corpus lexicon to the lemon model. The conversion process includes an intermediate phase in which the proprietary format is transformed into the CONLL-U (Conference on Natural Language Learning) format. A technological application was therefore developed to automatically convert linguistic annotations in CoNLL format into linked data (OntoLex – Lemon model). This program will be applied to the annotated Somali corpus to structure the terms of interest into a computational terminology.

Furthermore, to allow an initial annotation phase of the project’s reference corpora, a temporary user interface has been developed, also useful for supporting the process of defining user requirements that will lead to the development of the definitive annotation tool. The technologies in use are consistent with what is expected for the subsequent development phases. The tool allows scholars to annotate a first group of Jewish texts with lexical information, both in their original language and in their Italian translation.