The ARRAU Corpus of Anaphoric Information

Item Name: The ARRAU Corpus of Anaphoric Information
Author(s): Massimo Poesio, Ron Artstein, Olga Uryupina, Kepa Rodriguez, Francesca Delogu, Antonella Bristot, Janet Hitzeman
LDC Catalog No.: LDC2013T22
ISBN: 1-58563-663-0
ISLRN: 462-157-606-044-8
Release Date: December 16, 2013
Member Year(s): 2013
DCMI Type(s): Text
Data Source(s): meeting speech, microphone conversation, newswire, transcribed speech
Application(s): discourse analysis, information detection, information extraction, parsing, tagging
Language(s): English
Language ID(s): eng
License(s): LDC User Agreement for Non-Members
Online Documentation: LDC2013T22 Documents
Licensing Instructions: Subscription & Standard Members, and Non-Members
Citation: Poesio, Massimo, et al. The ARRAU Corpus of Anaphoric Information LDC2013T22. Web Download. Philadelphia: Linguistic Data Consortium, 2013.
Related Works: View


The ARRAU (Anaphora Resolution and Underspecification) Corpus of Anaphoric Information was developed by the University of Essex and the University of Trento. It contains annotations of multi-genre English texts for anaphoric relations with information about agreement and explicit representation of multiple antecedents for ambiguous anaphoric expressions and discourse antecedents for expressions which refer to abstract entities such as events, actions and plans.

The source texts in this release include task-oriented dialogues from the TRAINS-91 and TRAINS-93 corpora (the latter released through LDC, TRAINS Spoken Dialog Corpus LDC95S25), narratives from the English Pear Stories (a collection of narratives by subjects who watched a film and then recounted its contents), articles from the Wall Street Journal portions of the Penn Treebank (Treebank-2 LDC95T7) and the RST Discourse Treebank LDC2002T07, and the Vieira/Poesio Corpus which consists of training and test files from Treebank-2 and RST Discourse Treebank.


The texts were annotated using the ARRAU guidelines which treat all noun phrases (NPs) as markables. Different semantic roles are recognized by distinguishing between referring expressions (that update or refer to a discourse model), and non-referring ones (including expletives, predicative expressions, quantifiers, and coordination). A variety of linguistic features were also annotated, including morphosyntactic agreement, grammatical function, semantic type (person, animate, concrete, action, time, other abstract) and genericity. The annotation was carried out using the MMAX2 annotation tool which allows text units to be marked at different levels.

The files in MMAX format have been organized so that they can be visualized using the MMAX2 tool or directly used as input/output for the BART toolkit which performs automatic coreference resolution including all necessary preprocessing steps.


Please view the following samplesL


None at this time.

Available Media

View Fees

Login for the applicable fee