Restrictive relative constructions in Duala

This dissertation investigates the morphosyntax of restrictive relative constructions in Duala, a Bantu language distinguished by its unique morphological, phonological, and syntactic features. Motivated by ongoing debates in generative syntax—specifically the modification and connectivity problems—the study examines how relative clauses in Duala attach to their head noun phrases and how a single head noun can simultaneously serve functions in both the matrix and relative clauses. Data were collected through multiple phases of structured interactions with native Duala speakers, including in-person fieldwork and remote elicitation during the Covid-19 pandemic. These interactions, conducted primarily in French, involved grammaticality judgments and translation tasks that provided a robust empirical basis for the analysis. Utilizing the Minimalist Program as the theoretical framework, the research demonstrates that the subject marker in Duala functions not as a thematic subject but as an agreement morpheme that evidences phi-feature correspondence between the thematic subject and the verb. Detailed movement tests—including binding, reconstruction, and Duala-specific diagnostics such as the analysis of the -nɔ́ particle—support the argument that the connectivity problem is best addressed by a Head Raising Analysis rather than a Matching Analysis, particularly given the presence of Condition-C effects in relative constructions. The study also offers a typological account of Duala’s relative clauses by contrasting object and subject restrictive relatives. It identifies two primary relative elements: an optional element with ambiguous categorization (either a relative pronoun or complementizer) and a demonstrative-based element that is mandatorily present in subject relatives. The evidence indicates that Duala employs a gap strategy for relativizing core grammatical roles while requiring additional morphological modifications for more peripheral elements, as seen in the relativization of possessors. Further analysis of A'-dependencies reveals clear asymmetries. The data distinguish topicalization—resulting from base-generation—from other A'-dependencies that arise via movement, as well as a differential distribution of the -nɔ́ particle. Moreover, the contrasting behavior between subject focus and subject WH/relative constructions is attributed to the function of demonstrative elements operating as PF-repairs when subject displacement is not overt. The findings contribute both descriptively and theoretically. Empirical insights from Duala call for a reexamination of established syntactic models in Bantu languages and highlight the limitations of models predominantly based on Indo-European data. By providing new descriptive data and challenging existing assumptions about subject marking and relativization, this research advances our understanding of language universals and the cognitive mechanisms underlying human language. Future research is encouraged to extend these findings by exploring additional relative constructions in Duala, such as appositives, free relatives, and adverbial relatives, as well as conducting comparative studies with other Bantu languages.