Annotations are contained in .tsv files, which have two components (quote annotations and speaker attribution annotations); each type has its own structure:
| Label | Quote ID | Start sentence ID | Start token ID (within sentence) | End sentence ID | End token ID (within sentence) | Quotation |
|---|---|---|---|---|---|---|
| QUOTE | Q342 | 54 | 0 | 54 | 13 | “ Of course , we ’ll take over your furniture , mother , ” |
| Label | Quote ID | Speaker ID |
|---|---|---|
| ATTRIB | Q342 | Winnie_Verloc-3 |
Sentence IDs correspond to the row of the sentence in the corresponding .txt file; token IDs correspond to index of the token within that sentence.