Emergent projecting constructions

The present analysis is grounded in a view of grammar emerging in interaction and coming into being through mundane language use. By analyzing Hebrew interactional data, I outline the continua of synchronic usage from literal constructions involving the verb yada (‘know’) to three projecting constructions of the discourse marker variety. The study furthers our understanding of how projecting constructions are sedimentations of interactional practices. I combine interactional linguistics with grammaticization studies to show phonological, morphological, syntactic, pragmatic, and prosodic evidence in support of different grammaticization paths from matrix clause to discourse marker. Diachronic evidence from Mishnaic and Medieval Hebrew is added as further support for the paths suggested. The previously debated issue of complement-taking predicates is thus situated in another language and in a wider context, showing that different constructions employing the same matrix verb may follow different grammaticization paths. The study continues along an existing path of investigation into the issue of complement-taking constructions viewed as grammatically superordinate vs. as attachments to essentially monoclausal utterances. I argue for a more differentiated analysis recognizing both types of usage.