Coconstructions in spoken data: UD annotation guidelines and first results

TL;DR

This paper introduces annotation guidelines for cross-turn syntactic dependencies in spoken language, proposing two representations and new distinctions to improve spoken language treebanks within the UD framework.

Contribution

It presents novel annotation guidelines and two representations for cross-turn dependencies, enhancing spoken language syntactic annotation within the Universal Dependencies framework.

Findings

Proposed speaker-based and dependency-based representations for cross-turn dependencies.

Introduced new distinctions between reformulations, repairs, and unfinished elements.

Enhanced spoken language treebanks with detailed annotation guidelines.

Abstract

The paper proposes annotation guidelines for syntactic dependencies that span across speaker turns - including collaborative coconstructions proper, wh-question answers, and backchannels - in spoken language treebanks within the Universal Dependencies framework. Two representations are proposed: a speaker-based representation following the segmentation into speech turns, and a dependency-based representation with dependencies across speech turns. New propositions are also put forward to distinguish between reformulations and repairs, and to promote elements in unfinished phrases.