The AnnCor CHILDES Treebank

The AnnCor CHILDES Treebank Jan Odijk,Alexis Dimitriadis,Martijn van der Klis,Marjo van Koppen,Meie Otten,Remco van der Veen Utrecht University Trans 10 3512 JK Utrecht {j.odijk, a.dimitriadis, m.h.vanderklis, j.m.vankoppen, m.otten1, r.p.vanderveen}@uu.nl Abstract This paper (1) presents the first partially manually verified treebank for Dutch CHILDES corpora, the AnnCor CHILDES Treebank; (2) argues explicitly that it is useful to assign adult grammar syntactic structures to utterances of children who are still in the process of acquiring the language; (3) argues that human annotation and automatic checks on this annotation must go hand in hand; (4) argues that explicit annotation guidelines and conventions must be developed and adhered to and emphasises consistency of the annotations as an important desirable property for annotations. It also describes the tools used for annotation and automated checks on edited syntactic structures, as well as extensions to an existing treebank query application (GrETEL) and the multiple formats in which the resources will be made available. Keywords: treebank, Dutch, CHILDES, GrETEL, treebank querying 1. Introduction We describe the approach to the development of the Ann- Cor CHILDES Treebank, a treebank for Dutch CHILDES corpora. The whole treebank has been automatically gen- erated by the Alpino parser, but a part of this corpus has also been manually checked and, where needed, corrected. The AnnCor treebank is being created in the Utrecht Uni- versity AnnCor project, which we describe in section 2..1 We describe the syntactic annotation in this treebank in sec- tion 3.: we discuss some methodological issues with regard to the annotation (section 3.1.), the cleaning of CHILDES utterances (section 3.2.), the tool we use to inspect and edit syntactic structures (section 3.3.), annotation conventions we developed (section 3.4.), and checks on edited syntactic structures (section 3.5.). We have extended an existing tree- bank query application, and the extensions are described in section 4.. In section 5. we describe the formats in which the resources will be made available. In section 6. we dis- cuss related work, and we end with conclusions and plans and suggestions for future work in section 7.. This paper (1) presents the AnnCor CHILDES Treebank, the first partially manually verified treebank for Dutch CHILDES corpora; (2) argues explicitly that it is useful to assign adult grammar syntactic structures to utterances of children who are still in the process of acquiring the language; (3) argues that human annotation and automatic checks on this annotation must go hand in hand to com- bine the human’s intelligence and the software’s rigor; (4) argues that explicit annotation guidelines and conventions must be developed and adhered to. Requiring consistency of the annotations is important and should be an essential ingredient of such annotation conventions. 1This paper contains many hyperlinks hidden under terms and acronyms. The presence of a hyperlink is visible in digital ver- sions of the paper but may be badly visible or invisible in printed versions of the paper. 2. The AnnCor Project The AnnCor project2 is an Utrecht University internal re- search infrastructure project that aims to …

Content truncated to 500 words