ChiHwesa Verbal Constructions

This dataset constitutes a structured repository designed to examine the semantic compatibility of verbal extensions in ChiHwesa, with particular focus on the reciprocal (-an-), passive (-iw-/-w-/-h-), and stative (-ik-/-ek-) extensions. The underlying research hypothesis is that the distribution of these extensions is semantically constrained, such that only verbs with compatible argument structure and lexical semantics can productively undergo specific derivational processes. In particular, reciprocal constructions are expected to occur primarily with transitive and interactional verbs, passive constructions with verbs that license a patient/theme argument, and stative constructions with verbs that encode potentiality or resultative states. The dataset comprises a systematically organised list of ChiHwesa verb stems presented alongside their derived forms and corresponding English glosses. Each entry includes a base verb, its meaning, and its reciprocal, passive, and stative forms where attested. Importantly, the dataset also records gaps where certain derivations are not possible, providing critical evidence for semantic restrictions. The verbs included represent a wide range of semantic classes, including transitive, intransitive, cognitive, physiological, and social interaction verbs. Data were collected through a combination of native speaker intuition, structured elicitation, and cross-speaker validation. Informants were asked to generate and evaluate derived verb forms, and all entries were verified for consistency and acceptability. Additional support was drawn from an existing ChiHwesa lexical database. This multi-method approach ensures both reliability and linguistic authenticity. The data show that reciprocal formation is restricted to verbs denoting mutual or bidirectional actions, while verbs lacking inherent interaction (e.g., ‘walk’ or ‘urinate’) typically do not permit reciprocal derivation. In contrast, passive constructions are highly productive across verb classes, although some yield marginal or context-dependent interpretations. Stative constructions are also widely attested and primarily encode potentiality (‘able to be X-ed’) or resultant states, particularly with change-of-state verbs. These findings support the interpretation that verbal extensions in ChiHwesa are governed by semantic and argument structure constraints. The dataset is therefore valuable for theoretical analysis within frameworks such as Lexical Mapping Theory, as well as for comparative Bantu studies, language documentation, and computational applications. It can be used as both a descriptive resource and a basis for further hypothesis testing on the interaction between morphology, syntax, and semantics in under-documented languages.