The BMDS Dataset contains texts, annotations, and metadeta for detective stories published between 1890–1920.

• The stories are provided in plain text
• Metadata is provided for stories and authors
• Each story is annotated for 61 categories related ot types of crimes, clues, and evidence represented

Story-level annotations and metadata can be found in the BMDS_story_annotations.csv file. This file includes a unique Story Code for each story, which corresponds to the name of the plain text version of the story, which can be found in the texts/ folder. It also includes an Author Code for each story (two in the case of co-authored stories), which correspond to the author-level metadata in the BMDS_author_metadata.csv file.

The full text of the annotation guidelines that were used to prepare this dataset can be found at this link. These guidelines present detailed descriptions of all terms and categories used in the dataset.

The BMDS_story_annotations.csv contains the following columns, all of which are defined the Annotation Guidelines linked above.

• Story Code: corresponds to full text file ([Story Code].txt) in texts/ subfolder.

• Annotator #1 Code
• Annotator #2 Code
• Story Title
• Plot Summary
• Content warnings
• Author Code
• Second author code (if applicable)
• Investigation-reveal order
• Reveal border sentence
• Name of Detective #1
• Detective #1 Gender
• Detective #1 Role
• Name of Detective #2
• Detective #2 Gender
• Detective #2 Role
• Number of detectives if more than 2
• Name of Assistant #1
• Assistant #1 Gender
• Assistant #1 Role
• Name of Assistant #2
• Assistant #2 Gender
• Assistant #2 Role
• Number of assistants if more than 2
• Number of victims of gender Male
• Number of victims of gender Female
• Number of victims of gender Non-binary
• Number of victims of gender Unknown
• Number of victims who are corporate entities
• Number of culprits of gender Male
• Number of culprits of gender Female
• Number of culprits of gender Non-binary
• Number of culprits of gender Unknown
• Culprits introduced during reveal?
• Main culprit independence
• Initiator of investigation
• Role of police
• Focus on crime or quasi-crime
• Crime trajectory
• Occurrence of crime or mystery
• Types of qrimes
• Motives
• Means (murder only)
• Sufficient clues to guess?
• Sufficient clues to solve?
• Correct annotator guess?
• Types of clues
• Essential clue
• Type of essential clue
• Most salient clue
• Type of most salient clue
• Red herring description
• Type of red herring
• Presence of planted or fabricated evidence
• Types of evidence made available
• Is the crime solved?
• Decision NOT to alert authorities?
• Satisfying narrative account?
• How satisfying as detective fiction?
• Recommend to friend?

• Date of First Publication (YYYY-MM-DD)
• Format of First Publication (Journal or Book)
• Name of Journal or Title of Book

Note: BMDS_story_annotations.csv contains non-ASCII Unicode characters such as →, “ and ”.

The BMDS_author_metadata.csv file contains the following columns:

• Author Code: Corresponds to Author Code field in BMDS_story_annotations.csv
• Surname(s)
• Given Name(s)
• Sex
• Date of Birth (YYYY-MM-DD)
• Country of Birth
• City/Town of Birth
• Date of Death (YYYY-MM-DD)
• Country of Death
• City/Town of Death
• Nationality/ies
• VIAF: link to VIAF for author

• Adam Hammond (Department of English, University of Toronto)
• Simon Stern (Faculty of Law & Department of English, University of Toronto)

Email: [email protected], [email protected]