De un sistema de anotación sintáctica para lengua escrita a otro para lengua oral

Resumo

In the course of the manual production of a syntactically analyzed spoken corpus sample, we explain how we adapt a pre-existing system of syntactic annotation for written language. We found that what requires modification in that system does not implies the addition of new tags or syntactic structures, but the incorporation of mechanisms for the treatment of what is proper to the oral language. Without starting from a prior delimitation of the units of analysis, we proceed recognizing them as we parse, regardless of the prosodic segmentation and identifying, among those collected in our annotation system, the most extensive syntactic structures possible; we also integrate into the analysis, by various means, the so-called oral dysfluencies (repetitions, truncations and reformulations), and, reusing the labels that we already used in the annotation for this in written language, we account for the greater presence in the orality of markers, connectors and vocatives.