BECauSE 1.0

Intra-sentence Polarity Documented, Not Converted No Strength Signal

BECauSE 1.0 is the original release of the BECauSE corpus, annotating causal constructions across three source corpora. It predates BECauSE 2.0, the actively-maintained ~20%-larger release this project already has.

Each causal instance carries a Degree = FACILITATE | INHIBIT attribute — INHIBIT marking causation that prevents or hinders its effect, functionally a countercausal signal — a scheme that originates in this 1.0 release itself and is carried over unchanged into BECauSE 2.0; it is not yet converted at all, since v1.0's conversion is fully LDC-blocked (see below).

Blocked — mostly LDC-gated, remaining portion too small

Of v1.0's three source corpora, only CongressionalHearings (3 documents) ships its raw text directly in the release. The other two — NYT (59 documents, New York Times Annotated Corpus) and PTB (47 documents, Penn Treebank sections 2–23) — require separate LDC subscriptions this project does not have; only their annotations, not the underlying text, are in the repo. A conversion limited to the 3 free documents would be far too small to be a meaningful dataset and would not be representative of this corpus's cited size. Not attempted — see resources/datasets/BECauSEv1/conversion_script.py for the full trail.

Overview

Domain	News / Government
Year	2015
Sentences	—
Granularity	Intra-sentence

Tasks & Splits

Task	Train	Test

Citation

@inproceedings{dunietz:2015,
  title = {Annotating {{Causal Language Using Corpus Lexicography}} of {{Constructions}}},
  booktitle = {Proceedings of {{The}} 9th {{Linguistic Annotation Workshop}}, {{LAW}}@{{NAACL-HLT}} 2015, {{June}} 5, 2015, {{Denver}}, {{Colorado}}, {{USA}}},
  author = {Dunietz, Jesse and Levin, Lori S. and Carbonell, Jaime G.},
  editor = {Meyers, Adam and Rehbein, Ines and Zinsmeister, Heike},
  year = 2015,
  pages = {188--196},
  publisher = {The Association for Computer Linguistics},
  doi = {10.3115/V1/W15-1622},
  url = {https://doi.org/10.3115/v1/w15-1622},
  urldate = {2025-05-08}
}