The project has the urgent task to create and archive a repository of speech samples of various length and genres in Bayso and Haro, along grammatical sketches, dictionaries and sociolinguistic and anthropological profiles. The material is recorded in audio and video. The documentation activities are multidisciplinary and are based on the collaboration of linguists and anthropologists. All the team members collect, transcribe and translate the recorded speech with the collaboration of local consultants. Linguists are in charge of grammatical annotation and the creation of grammatical sketches, dictionaries and sociolinguistic profiles. The anthropologists add cultural information about the texts and create anthropologic profiles. Data collection is conducted both in the Gidiccho island and in the village of Alge. Visits are also paid to the other villages where some of the Bayso and Haro live. Recordings are made on digital support with the use of recorders providing uncompressed (.wav) sound files. Speech is provided by a wide range of speakers of different age, gender, and social status. A more conservative form of the languages is expected from people who spend most of their time on the Gidiccho island. Here they also preserve practices that tend to disappear in the mainland villages. Knowledgeable elders will be the reference speakers for the description of ancient and lost practices, while younger and educated speakers assist in transcription and translation. Other subjects involved are anthropology University students that have field experience in the area and other people who know the area and can provide practical assistance. The texts are transcribed and translated on paper in the field. An agreed Latin orthographic system is used since these languages have no standard script. The working language is Amharic. Most of the team member speak Amharic, but local consultants who know English will also be used as intermediate translators in case of need. The transcription and the translation found in field notes are checked and typed in in an ELAN file. From the transcription, the most interesting texts are annotated grammatically still in ELAN. In particular, the ELAN-Corpa version of this program will be used as it has an additional function of semi-automatic annotation. The culturally most relevant recordings are enriched with anthropological notes.
