This repository contains the data set created for the following research papers:

Savelka, Jaromir, and Kevin D. Ashley. "Discovering Explanatory Sentences in Legal Case Decisions Using Pre-trained Language Models." Findings of the Association for Computational Linguistics: EMNLP 2021. 2021.

Jaromir Savelka, Huihui Xu, and Kevin D. Ashley. 2019. Improving Sentence Retrieval from Case Law for Statutory Interpretation. In Seventeenth International Conference on Artificial Intelligence and Law (ICAIL ’19), June 17–21, 2019, Montreal, QC, Canada, Floris Bex (Ed.). ACM, New York, NY, USA, 10 pages. https://doi.org/10.1145/3322640.3326736

Given a statutory provision, user's interest in the meaning of a phrase from the provision, and a list of sentences we would like to rank more highly the sentences that elaborate upon the meaning of the statutory phrase of interest, such as:

• definitional sentences (e.g., a sentence that provides a test for when the phrase applies)
• sentences that state explicitly in a different way what the statutory phrase means or state what it does not mean
• sentences that provide an example, instance, or counterexample of the phrase
• sentences that show how a court determines whether something is such an example, instance, or counterexample.

For this corpus we selected fourty two terms from different provisions of the United States Code.

For each term we have collected a set of sentences by extracting all the sentences mentioning the term from the court decisions retrieved from the Caselaw access project data.

In total the corpus consists of 26,959 sentences.

The sentences are classified into four categories according to their usefulness for the interpretation:

• high value - sentence intended to define or elaborate on the meaning of the term
• certain value - sentence that provides grounds to elaborate on the term's meaning
• potential value - sentence that provides additional information beyond what is known from the provision the term comes from
• no value - no additional information over what is known from the provision

See Annotation guidelines for additional details.

Each zip file contains data related to one of the fourty two queries. There are four files in total containing the texts of different granularity. These allow to replicate experiments reported in the paper cited above.

• case
  • original_id - case id from Caselaw access project
  • name
  • short_name
  • date
  • official_date
  • official citation
  • alternate_citations
  • court
  • short_court - court abbreviation
  • jurisdiction
  • short_jurisdiction - jurisdiction abbreviation
  • attorneys
  • parties
  • judges
  • text
• opinion
  • case_id - pointer to the case the opinion belongs to
  • author
  • type - e.g., concurrence, dissent
  • position - position of the opinion within the case
  • text
• paragraph
  • case_id - pointer to the case the opinion belongs to
  • opinion_id - pointer to the opinion the paragraph belongs to
  • position - position of the paragraph within the opinion
  • text
• sentence
  • case_id - pointer to the case the sentence belongs to
  • opinion_id - pointer to the opinion the sentence belongs to
  • paragraph_id - pointer to the paragraph the sentence belongs to
  • position - position of the sentence within the paragraph
  • text
  • label - human-created gold label of the sentence value

For use of the data we kindly ask you to provide the two following attributions:

Savelka, Jaromir, and Kevin D. Ashley. "Discovering Explanatory Sentences in Legal Case Decisions Using Pre-trained Language Models." Findings of the Association for Computational Linguistics: EMNLP 2021. 2021.

The President and Fellows of Harvard University, Caselaw access project, Caselaw access project, 2018.