EventCausality

Intra-sentence No Polarity Signal No Strength Signal

EventCausality provides minimally-supervised annotations of causal relations between events collected from web text, using a small set of seed patterns to bootstrap annotation.

No direct source available — data quality unresolved

Unlike most datasets in this toolkit, EventCausality could not be moved off the CREST aggregation to a direct source. The original release (Do, Chan & Roth, EMNLP 2011 — 580 manually-annotated causal links on 25 CNN articles) was never publicly archived anywhere reachable: the paper only promises a future release via a CogComp resource page that is now dead. TCR covers the same 25 source documents but is a different, later re-annotation (different event set, different label scheme, 172 vs. 580 links) — using it here would silently conflate two distinct datasets under one citation, so this repo does not do that. The data below still comes from CREST's own aggregation, which has a confirmed corruption/completeness bug (0 train rows) and a confirmed character-offset misalignment affecting causal-candidate-extraction and causality-identification alike (~21% of spans are fully out-of-bounds, now dropped rather than silently shipped) — treat results on this dataset with those caveats in mind.

Overview

Domain	Web
Year	2011
Sentences	583
Granularity	Intra-sentence

Tasks & Splits

Task	Train	Test
Causality Detection	—	—
Causal Event Candidate Extraction	—	—
Causality Identification	—	—
Causality Extraction	—	—

Data Explorer

Causality Detection

Causal Event Candidate Extraction

Causality Identification

Citation

@inproceedings{do:2011,
  title     = {Minimally {{Supervised Event Causality Identification}}},
  booktitle = {Proceedings of the 2011 {{Conference}} on {{Empirical Methods}} in {{Natural Language Processing}}, {{EMNLP}} 2011, 27-31 {{July}} 2011, {{John McIntyre Conference Centre}}, {{Edinburgh}}, {{UK}}, {{A}} Meeting of {{SIGDAT}}, a {{Special Interest Group}} of the {{ACL}}},
  author    = {Do, Quang and Chan, Yee Seng and Roth, Dan},
  year      = 2011,
  pages     = {294--303},
  publisher = {ACL},
  urldate   = {2026-01-05}
}