Anaphora refers to a phenomenon in which one linguistic form depends on another for its reference, such as the way in which the English reflexive "himself" refers back to "John" in "John likes himself". Theoretical work on anaphora in natural language has been a major engine for innovation in linguistics for over 30 years because of deep cross-linguistic similarities. Such similarities provide important evidence for the existence of a shared human language faculty and provide important details about the workings of that faculty. At the same time, languages appear to differ a great deal in their particular patterns of anaphora and such differences are also important in linguistic theorizing. The goal of this project is to expand our knowledge of empirical patterns of anaphora in the non-colonial languages of Africa through collection and in-depth analysis of data from these languages. In earlier work supported by the NSF, the project performed a number of detailed case studies of anaphora in selected African languages. That work has already uncovered new patterns of anaphora informing the nature of transitivity, the nature of clause boundaries, the competition of anaphoric forms, and some unexpected interpretive subtleties. The research methodology, to be continued in this project, consists in enlisting native speaker linguists as language consultants to fill out a detailed questionnaire made available over the internet. Both the questionnaire responses and language-related materials are made available at the project's web site and in a specially designed database also developed with NSF funding that permits fine-grained comparative work (www.africananaphora.rutgers.edu).
The data from this project will be of interest to theoretical linguists working in a variety of frameworks. The comprehensiveness of the database's case files will contribute to the goal of advancing linguistic research on anaphora. African languages are particularly valuable in this goal: of the estimated 6000 languages of the world, approximately 2000 are spoken in Africa, and Africa is unique in having so many speakers of national and minority languages with sufficient academic training to participate in the project. Collaboration between the project staff and native speaker linguists of Africa permits knowledge to flow in both directions, building the community of scholars that focus on fundamental questions about humans' linguistic abilities. The research also develops a methodology that can be a model for comparative research into other empirical patterns in the world's languages. An ultimate goal of this project is to involve theorists and native speaker consultants in new investigations of empirical patterns beyond anaphora, using the project's existing infrastructure, including website, database, consultant network, and project staff.
The Afranaph Project has developed a new platform for research into African Languages, most of them understudied and some of them endangered, it has developed a research community around the new digital information resources that have been developed, it has provided a model for linguistic field research in the age of teh web and the internet, both in terms of public access and technique, and it has been a vehicle for educating the next generation of researchers both in the US and Africa. The data collected from over 40 languages, 30 of which are already reported on the Afranaph Website and Database, continues to expand as our project is ongoing based on new funding. As our fine-greaned understanding of a variety of newly investigated languages emerges, new opportunities are created for understanding the human language faculty and its diverse manifestations. Research results obtained from our methods and the permanent resources created by the project have already influenced theoretical discussion in a variety of areas, particularly in the domain of anaphora, but as much new work has only just been published or will appear in the next few years, the ultimate impact of the Afranaph Project on the direction of linguistic theory can nly be evaluated at a later date.
