The Data

The information to be analysed by this project is summarized as follows:

Genealogies of the inhabitants of the village of Somié -

Census data from the village of Somié

- 1990

- 1995/6

Baptismal registers of Protestant and Catholic Churches since 1945.

State birth and death registers since 1955.

Census from Atta 1954

During the pilot project we will establish a suitable database structure, and enter the information collected in the course of previous research and fieldwork among the Mambila, including the results of censuses which I undertook in Somié in 1990 and 1995. In addition, the returns of a census conducted in 1954 in the neighbouring village of Atta, made available to me by Jean Hurault, will be recorded.

Each type of data will be entered separately to maintain its integrity, so that the information will be of use to subsequent researchers. At the point of data entry no assumptions will be made about the identity of individuals with similar or identical names.

Links between the files, and between the individuals in the files, will be established subsequent to data entry. Genealogical data which I have collected over the last ten years will be applied to help establish the links between the individuals. Since this data has been collected and checked throughout my fieldwork in Somié it is very high quality. Techniques will be applied to overcome the problems inherent in the study of historical data, including that of variant spellings of names. Specifically, existing soundex programs already written by Dr Bagg will be modified in the light of Mambila phonology to effect name matching. The method favoured at present is a coding system based in principle upon the Russell soundex but using an all numeric code and taking account of relative positioning for crucial letters - e.g.. consonant plus a following vowel. A look-up table of name equivalents and abbreviations (e.g. Mbeya : Mia, Ndinuaga : Nuaga) will be needed with this system. Other systems, such as that developed by Lars Nygaard (1992), are being considered.

Mambila genealogies tend to be wide and shallow. The greatest depth, seven generations, is exceptional, four generations being typical. The time scale is thus comparable with that of the baptismal records (adults born in the 1920s-1940s, in first decades of evangelisation, were baptised).

The demographic analysis will use techniques for dealing with incomplete evidence as discussed by authors in Fetter (ed.) and the Cambridge Group for the History of Population and Social Structure (e.g. Smith and Oeppen 1993).



Head of Proposal - Evaluation - VIMS