Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Coconstructions in spoken data: UD annotation guidelines and first results

Ludovica Pannitto, Sylvain Kahane, Kaja Dobrovoljc, Elena Battaglia, Bruno Guillaume, Caterina Mauri, Eleonora Zucchini

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.

Coconstructions in spoken data: UD annotation guidelines and first results

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.

Paper Structure

This paper contains 18 sections, 7 figures.

Table of Contents

  1. Introduction
  2. The units of spoken UD treebanks
  3. A typology of coconstructions
  4. Coconstruction proper
  5. A. Speaker 2 realizes an already projected dependency.
  6. B. Speaker 2 adds a new dependency.
  7. Wh-questions
  8. Backchannelling
  9. Proposed guidelines
  10. Completion of an open dependency slot
  11. Stacking on an already realized dependency slot
  12. Automatic conversion between speaker-based and dependency-based
  13. Coconstructions in existing UD treebanks
  14. French (Rhapsodie)
  15. Italian (KIParlaForest)
  16. ...and 3 more sections

Figures (7)

  • Figure 1: An Italian example of a conj:reform coconstruction and a backchannel. The example is taken from KIParla, BOA3017. Tr. en. 'apostrophe', 'yes, apostrophe 'mbare', 'mh'. 'mbare is a dialectal form for compare, 'mate'.
  • Figure 2: Italian example (\ref{['ex:mestiere']}) with promotion (NB: the sentence is shortened, for space reasons). The first line shows speaker-based annotation, the second line shows the intermediate graph representation of the dependency-based view and line 3 shows final UD well-formed tree for the dependency-based view.
  • Figure 3: Original transcription of Example (\ref{['ex:mestiere']}) in the KIParla corpus.
  • Figure 4: English example with lifting. Line 1: speaker-based annotation; Line 2: intermediate representation of the dependency-based view; Line 3: final UD tree for the dependency-based view
  • Figure 5: Slovenian example of promotion ‘[…] that he once had with her a pleasant … experience.’
  • ...and 2 more figures