A core purpose of human language is to convey information about events. We refer to the people and things involved in these events in every clause we create. Crucially, in conveying information we follow the grammatical rules of our language: nouns cannot simply be put in any form or position, but they need to be licensed by the grammar. Such nominal licensing has hitherto been connected to case (nominative, accusative) and grammatical roles (subject, object). However, the Bantu languages seem to organise nominal licensing along other properties. In these languages it is much more important how the information flow is structured; this concerns for example whether a referent is already active in the hearer’s mind or not, or whether the speaker wants to contrast a referent with potential alternatives (e.g. ‘I ate rice not plantains’). The information structure has a fundamental effect on the grammar of Bantu languages, but quite how this affects nominal licensing is as yet unknown.

This project aims to systematically investigate the influence of information structure on nominal licensing in a subset of Bantu languages. Going beyond the simple notions of ‘topic’ and ‘focus’, our research diagnoses multiple aspects of information structure for each language (activation states, unexpectedness, implied or encoded contrast with alternatives etc.). The new data will contribute to a better understanding of the grammar of each of the languages, and also reveal cross-linguistic comparative patterns in the expression of information structure, for example to what extent subject and object markers on the verb indicate old information, or how word order expresses focus.

With the resulting comparative overview of syntax and information structure across Bantu we can test the theoretical hypothesis underlying the project: there is an alternative system of nominal licensing, in which the syntax is guided by ‘discourse salience licensing’ rather than ‘Case licensing’. The project thus aims to provide a more encompassing linguistic model that can accommodate the hitherto underappreciated ways in which grammatical organisation can vary.