Tabular dataset for "Syntactic-semantic capture of historical texts as a platform for source-critical analysis: telling the story of a premodern heresy trial with Computer-Assisted Semantic Text Modelling (CASTEMO)"

This tabular dataset supports this article: 
Robert L.J. Shaw, Katalin Suba, Tomáš Hampejs, David Zbíral, "Semantic-syntactic capture of historical texts as a platform for source-critical analysis: telling the story of medieval trial records with Computer-Assisted Semantic Text Modelling (CASTEMO)", Digital Scholarship in the Humanities (2026), https://doi.org/10.1093/llc/fqag016.

It contains data derived from the inquisitorial process against Bernard-Oth of Niort and his family (1234, or possibly 1235) as represented in Paris, Bibliothèque nationale de France, MS Doat 21, fols. 34r–50r. It provides details of the 113 witnesses (all male) whose testimony is extant within the record, and the responses they gave. It represents an analytical projection of the CASTEMO data collected from this text in InkVisitor, which can be viewed at https://inkvisitor-niort.dyn.cloud.e-infra.cz/.

The sheet contains the following columns:

• “witness_id": a number from 1 to 113 providing a unique identifier for each witness, following the order in which their depositions appear within the text. The missing witness deposition that can be inferred from the reference found in the deposition of Raimundus de Taranha (no. 30) is given the value "29a".

• "document_section": specifies the section of the document in which the witness deposition appears; "1" = 1st textual section, representing the 1st trial sitting (undated); "2" = 2nd section, representing the 3rd trial sitting (26 January); "3" = 3rd textual section, representing the 2nd trial sitting (1 February); "4" = 4th textual section, representing the 4th trial sitting (8 February).

• "witness_name": records the name of the witness.

• "occupation_or_office": records the name (in Latin) of the occupation or office held by the witness. "N/A" is recorded if no occupation or office is stated.

• "occupation_or_office_category": an analytical categorization of the witness's occupation or office; "clergyman" = secular clergyman; "religious" = a member of a religious order; "none" = lay or otherwise unknown status.

• "high_rank": a classification of social standing based on the witness's occupation of office: "1" = a high rank individual, typically with responsibilities over others within a hierarchy; otherwise "0".

• "associated_location_name": records the name (in Latin) of any location associated with the witness. In the few cases where witnesses had two locations associated with them, we have chosen just one based on the strength of association, preferring clearly stated locations of residence or employment over toponymic surnames. "N/A" is recorded if the witness has no associated location. 

• "associated_location_type": states the nature of relationship between the associated location and the witness; "office" or “occupation" = place of work; "location of residence" = clearly stated place of residence; "human spatial reference point" = undefined association derived from toponymic surname. "N/A" is recorded if the witness has no associated location. 

• "associated_location_lat" and "associated_location_long" record the latitude and longitude coordinates of any associated place, if successfully geocoded. "N/A" is recorded if the associated location was not successfully geocoded or if the witness has no associated location.

• "distance(kms)_to_Laurac": records the "as the crow flies" distance of any associated location from the town of Laurac, the centre of Niort family power in the Lauragais region.

• "charge_1_affirmation", "charge_2_affirmation", "charge_3_affirmation" and "charge_4_affirmation": an analytical categorization of the responses to each of the four charges against the members of the Niort family. If the charge was not affirmed by the witness, "unaffirmed" is recorded. In the cases where the charge was affirmed,  a three part string is recorded, each part separated by "_", e.g. "regular_own_all":

  • In the first string part, "regular" = affirmation not stated as reliant solely on hearsay; "hearsay" affirmation stated as reliant solely on hearsay.

  • In the second string part, "own" = affirmation stated in the witness's own deposition; "inherited" = affirmation inferred from textual reference to the deposition of another witness (e.g. X "said the same as" Y).

  • In the third string part, "all" = charge affirmed for all suspects within the Niort family; "some" = charge affirmed for only some members of the family.

• "testimony_beyond_charges": records whether the witness deposition contains details beyond direct responses to the charge; "own" = additional details stated in the witness's own deposition; "some" = additional details inferred from textual reference to the deposition of another witness (e.g. X "said the same as" Y). "N/A" is recorded if no additional details are stated or inferred from textual reference.

• "referenced_witness_id" and "referenced_witness_name": identify any textual reference to the deposition of another witness (e.g. X "said the same as" Y). "N/A" is recorded there is no such reference.