| This version of the corpus was made from the XML files. | |
| It is tagged by files (112), sections (363) and speakers (179). | |
| Files are named after the first word of the name of the case. For instance: the case "Air & Liquid Systems Corp. v. DeVries" corresponds to the file "Air". | |
| The sections correspond to the different oral arguments heard in one session. For instance, in "Air", there are three sections, because there were 3 arguments: | |
| 1. Oral argument of Shay Dvoretzky on behalf of the petitioners; | |
| 2. Oral argument of Thomas C. Goldstein on behalf of the respondents; | |
| 3. Rebuttal argument of Shay Dvoretzky on behalf of the petitioners. | |
| Each speaker was identified and their speaking turns tagged as being uttered by the corresponding speaker. It is therefore possible to study one speaker's speech in particular by creating a sub-corpus. | |
| When the corpus was imported into TXM, it was lemmatized and tagged in part of speech with Tree Tagger. | |
| We manually corrected part of speech tagging mistakes. We also corrected other errors, for instance speaker names when there were typos in their names or when two speakers had the same family name (we added the first letter of their first name to differentiate them). | |
| 2 sub-corpora are included in this version of the corpus: | |
| - sub-corpus "OA Amici" (oral arguments of the Friends of the Court) | |
| - sub-corpus "OA Avocats" (oral arguments of the attorneys, without those of the Friends of the Court). | |
| The "OA Avocats" sub-corpus is divided into two parts: judges' speaking turns (467 143 tokens) and lawyers' speaking turns (803 399 tokens). | |