# Entity Annotation Rules Instance - Section 3 of Convention v1.5.0
# Captures all entity annotation rules for the Gado2 annotation scheme
# Version 1.5.0 - Introduced double tagging support and THING entity type

annotation_scheme: "Gado2"
no_double_tagging: false  # CHANGED: Now allows overlapping annotations
context_sensitive: true
version: "1.5.0"
date_modified: "2025-11-12"

# General annotation policies
general_rules:
  - Double tagging is now ALLOWED and ENCOURAGED for semantic richness
  - The same span of text can receive multiple entity tags when it represents
    multiple semantic concepts (e.g., 'Kapitein Jonker' = BEING + DENOMINATION + DENOMINATION)
  - Double tagging applies to:
    * Titles/denominations within personal names
    * Places within personal descriptions
    * Objects within quantity expressions
    * Any semantically meaningful overlapping entities
  - The exact same word can be tagged in different ways depending on its context
  - When in doubt, check how words are tagged in previous ground truths
  - This convention distinguishes titles from general designations:
    * Titles are used/accepted by subjects (emic description)
    * General designations are imposed upon subjects (ethnic description)
  - Transcription status: Done → Final → Ground Truth (GT)
  - Nested and overlapping annotations should capture all semantic layers

double_tagging_patterns:
  - pattern: "Title + Personal Name"
    description: "Tag the full person name, plus denomination for title, plus surname separately if historically significant"
    examples:
      - text: "Kapitein Jonker"
        tags:
          - entity: "Kapitein Jonker"
            type: BEING
            span: [0, 15]
          - entity: "Kapitein"
            type: DENOMINATION
            span: [0, 8]
          - entity: "Jonker"
            type: DENOMINATION
            span: [9, 15]
            note: "Old spelling for Jonkheer (noble title/name)"

  - pattern: "Being + Place"
    description: "Tag the full person description plus embedded place reference"
    examples:
      - text: "Maria uit Bengaal"
        tags:
          - entity: "Maria uit Bengaal"
            type: BEING
            span: [0, 17]
          - entity: "Bengaal"
            type: PLACE
            span: [10, 17]

  - pattern: "Quantity + Thing"
    description: "Tag the full quantity expression plus the object being counted"
    examples:
      - text: "een lepel"
        tags:
          - entity: "een lepel"
            type: QUANTITY
            span: [0, 9]
          - entity: "lepel"
            type: THING
            span: [4, 9]
      - text: "twee kandelaars"
        tags:
          - entity: "twee kandelaars"
            type: QUANTITY
            span: [0, 15]
          - entity: "kandelaars"
            type: THING
            span: [5, 15]

entity_types:
  # 3.1 BEING
  - entity_type: BEING
    key: BEI
    label: BEING
    description: >-
      Personal names and specific references to persons (named or unnamed) or beings.
      Includes persons, animals with proper names, fictional characters,
      gods, spirits, saints and prophets. With double tagging enabled,
      titles and denominations within person names are also tagged separately.
    
    subcategories:
      - name: personal_names
        key: PNAME
        label: Personal Names
        description: Given name and surname, patronyms
        examples:
          - "Gouverneur-Generaal van Starkenborgh"
          - "President Soekarno"
          - "Sultan Muzaffar"
          - "Dr. Raden Soetomo"
          - "De Heer Balkenende"
          - "mr. dr. Beel"
          - "Tumenggung Sukapura"
          - "Officier Suksapura"
          - "Officier Tumenggung Sukapura"
          - "Lieutenant Winkelaar"
          - "Lieutenant Adolph Winkelaar"
          - "Mr. van Agt"
          - "Mej. den Uyl"
          - "Marquis De La Chetardie"
          - "Meneer Drees"
          - "gesaghebber Van Arrewijne"
          - "gewesen Sicacalsen Serlasker Abdul Nabij"
          - "princesse Gamoelamoe (vertaling van Ratu Gamulamu)"
          - "Engelschen vice admiraal Cornisch"
          - "Japara 's resident Lepeltak"
          - "Japaras resident Falck"
          - "resident Falck"
          - "vice admiraal Cornisch"
          - "Kapitein Jonker"  # ADDED: Also tag 'Kapitein' and 'Jonker' separately as DENOMINATION
          - "Maria uit Bengaal"  # ADDED: Also tag 'Bengaal' as PLACE

      - name: animal_names
        key: ANAME
        label: Animal Names
        description: Proper names of animals
        examples:
          - "Lassie"
          - "Golden Retriever"

      - name: fictional_characters
        key: FICT
        label: Fictional Characters
        description: Proper names of fictional characters and artifical beings
        examples:
          - "Kuifje"
          - "Sherlock Holmes"
          - "Hercule Poirot"
          - "{Suske} en {Wiske}"

      - name: religious_figures
        key: RELBEI
        label: Religious Figures
        description: Gods, spirits, saints and prophets
        examples:
          - "God"
          - "Jesus"
          - "Mohammed"
          - "Heilige Geest"
          - "Rama"

      - name: specific_references
        key: SPECBEI
        label: Specific References
        description: Specific references to a person without their name being mentioned. The denomination/title is also tagged separately with DENOMINATION
        examples:
          - "de Koning van Pruisen"
          - "de Koningin van Nederland"
          - "de Sultan van Ternate"
          - "de Minister van Buitenlandse Zaken"
          - "de Gouverneur-Generaal van Nederl.-Indië"
          - "de Koning"
          - "de Sultan"
          - "de Directeur"
          - "Tenno Heika"
          - "zijn Vrouw"
          - "haar slaaf"
          - "de geseijde slavin"
          - "de eerder genoemde leider"
          - "den Capiteijn"
          - "des Conincx"
          - "dit Jongetie"
          - "dat kind"
          - "Zijne Majesteit"
          - "haaren Coning"
          - "dezen Prins"
          - "dien Coning"
          - "onsen Capiteijn"
          - "Bapak itu" # Indonesian
          - "gemelte gouvern:r"
          - "De slaaf"

    inclusion_rules:
      - rule_id: BEI_INC001
        description: Include titles with personal names - and tag titles separately as DENOMINATION
        conditions:
          - "Tag the full person name including titles"
          - "ALSO tag titles/denominations separately as DENOMINATION (e.g., Jonker/Jonkheer)"
        examples:
          - "Gouverneur-Generaal van Starkenborgh → BEING (full), Gouverneur-Generaal → DENOMINATION"
          - "President Soekarno → BEING (full), President → DENOMINATION"
          - "Sultan Muzaffar → BEING (full), Sultan → DENOMINATION"
          - "Kapitein Jonker → BEING (full), Kapitein → DENOMINATION, Jonker → DENOMINATION (noble name)"

      - rule_id: BER_INC002
        description: Include words between title and personal name
        conditions:
          - "The distance between a title and personal name can vary within a single phrase"
          - "If another word stands between the title and the personal name, it will be tagged along"
          - "Tag intermediate words separately if they have semantic meaning (e.g., place/denomination)"
        examples:
          - "Engelschen vice admiraal Cornisch → BEING (full), Engelschen vice admiraal → DENOMINATION"

      - rule_id: BEI_INC003
        description: Include articles and demonstratives with specific references
        conditions:
          - "The article/demonstrative/genitive case/possessive pronouns are tagged along"
        examples:
          - "de Koning van Pruisen → BEING, Pruisen → PLACE (separate)"
          - "zijn Vrouw"
          - "des Conincx"
          - "dit Jongetie"

      - rule_id: BEI_INC004
        description: Include backreferences
        conditions:
          - "References that point back to previously mentioned BEINGs"
        examples:
          - "gemelte gouvern:r"

      - rule_id: BEI_INC005
        description: Include embedded place references but also tag the places separately as PLACE
        conditions:
          - "When a BEING description includes a place reference (e.g., 'uit Bengaal'), tag the full BEING AND the place separately"
        examples:
          - "Maria uit Bengaal → BEING (full), Bengaal → PLACE"
          - "Jan van Amsterdam → BEING (full), Amsterdam → PLACE"

      - rule_id: BEI_INC006
        description: Include embedded associated organisationsbut also tag the organisations separately as ORGANISATION
        conditions:
          - "When a specific reference to a BEING includes an associated organisation or place (e.g., 'de Minister van Buitenlandse Zaken'), tag the full BEING AND the organisation/place separately"
        examples:
          - "De Minister van buitenlandse zaken → BEING (De Minister van buitenlandse zaken), ORGANISATION (buitenlandse zaken)"
    exclusion_rules:
      - rule_id: BEI_EXC001
        description: Do not tag abstract references
        rationale: >-
          Abstract references (often indicated by plural forms or indefinite
          articles) are not tagged, these often need to be tagged as 'denomination'
        examples:
          - "Een President"
          - "Een Premier"
          - "Een resident"
          - "Een generaal-majoor"
          - "Een Voorzitter"
          - "hij ontving de titel dr."
          - "ambassadeurs"
          - "Koninginnen"
          - "Koningen"
          - "draagt de titel Graaf"
          - "die persoon"
          - "de mannen"
          - "de vrouwen"
          - "die lieden"
          - "de kindere"
          - "Bapak" # Indonesian, without 'itu' (that) or other demonstrative
       

      - rule_id: BEI_EXC002
        description: Do not tag pronouns
        rationale: >-
          Pronouns are not tagged. They are consistent enough to be traced
          through regular expressions and do not require entity processing.
        examples:
          - "hij"
      - rule_id: BEI_EXC003
        description:  Tag general designations separately when they occur with complete personal names
        rationale:
          -  We tag both the full BEING AND the designation separately
        examples:
          - "geseijde slavin tanjong → BEING (tanjong), DENOMINATION (geseijde slavin)"
          - "de slaaf Anthonij → BEING (Anthonij or full), DENOMINATION (slaaf)"
          - "de Javaanschen Cap:n Soeta Wangsa → BEING (full), DENOMINATION (Javaanschen Cap:n)"

  # 3.2 PLACE
  - entity_type: PLACE
    key: PLC
    label: Place
    description: >-
      Geographic locations including streets, cities, provinces, countries,
      continents, infrastructure, landforms, public spaces, buildings, and
      astronomical objects. With double tagging enabled, places embedded in
      BEING descriptions, denominations, or other entities are also tagged.

    subcategories:
      - name: street_names_addresses
        key: SNAME
        label: Street Names and Addresses
        description: Street names and addresses
        examples:
          - "Laan van Meerdervoort"

      - name: trajectories_directions
        key: TDIR
        label: Trajectories and Directions
        description: Trajectories and wind directions
        examples:
          - "schoone wegh met seer schoone water plaatsen versien, streckende meest O:Z:O. en W:N:W. aen de noordt zijde"
          - "Oostzijde van Molenvliet"

      - name: locations
        key: LOC
        label: Locations
        description: Dorp, Stad, Provincie, Land, Continent, Bisdom, Zones
        examples:
          - "Sectie [0-9, I-V]"
          - "Bengaal"  # ADDED: Can appear within BEING descriptions
          - "Indonesia"  # ADDED: Can appear as adjective base in denominations

      - name: infrastructure
        key: INFRA
        label: Infrastructure
        description: Brug, Haven, Dam
        examples:
          - "de haven van Rotterdam"
          - "de brug over de Amstel"

      - name: landform
        key: LANDF
        label: Landform
        description: Berg, Gebergte, Bos, Rivier, Bron, Rijstvelden, Vallei, Tuin, Natuurreservaat, Nationaal park, Plantage, Strand
        examples:
          - "de rivier de Maas"
          - "het bos bij Arnhem"
          - "de berg Merapi"
          - "de plantage Deli"

      - name: public_space
        key: PUBSP
        label: Public Space
        description: >-
          Plein, Veld, Theater, Museum, School, Markt, Vliegtuig, Station,
          Zwembad, Ziekenhuis, Sportveld, Bioscoop, Tentoonstelling, Campus,
          Lanceerplatform, Club, Huis, Universiteit, Bibliotheek, Bidruimte,
          Medisch centrum, Parkeergarage, Speeltuin, Grafplaats, Bedrijvenpark
        examples:
          - "het museum van Rotterdam"
          - "de school aan de Laan van Meerdervoort"
          - "het ziekenhuis St. Antonius"

      - name: companies_as_places
        key: COMPPLC
        label: Companies as Places
        description: >-
          Companies only if contextualised as places (otherwise organisation):
          Apotheek, Bar, Restaurant, Eettent, Depot, Hotel, Hostel, Fabriek,
          Nachtclub, Muziekpodium
        examples:
          - "het restaurant De Kromme Watergang"
          - "het hotel Grand Hotel Krasnapolsky"

      - name: buildings
        key: BLDG
        label: Buildings
        description: >-
          Huis, Wooncomplex, Klooster, Kleuterzaal, Flat, Kazerne, Fort,
          Verzorgingshuis, Winkelcentrum, Paleis
        examples:
          - "het paleis op de Dam"
          - "het fort aan de kust"

      - name: astronomical_objects
        key: ASTR
        label: Astronomical Objects
        description: Zon, Aarde, Maan, Planeten, Comets (often named after persons/dates)
        examples:
          - "de maan"
          - "de aarde"
          - "Mars"
          - "Halley"

      - name: coordinates
        key: COORD
        label: Coordinates
        description: Geographic coordinates
        examples:
          - "Z. breete van 35 — 31 ten langte 5 — 15"
          - "parabel van 36 — 50"
          - "3½ graed of 42 mijlen bewesten"
          - "10 graed noorder breedte"
          - "de evenaar"

    inclusion_rules:
      - rule_id: PLC_INC001
        description: Include relevant adjectives
        conditions:
          - "Directional and descriptive adjectives are tagged with place names"
        examples:
          - "Noord/Oost/Zuid/West(-kust)"
          - "Aziatische-Afrikaanse landen"
          - "Amerikaanse defensie-gebied"
          - "Asia Raya"
          - "Muang Thai"
          - "Indonesische eilandenrijk"
          - "Indonesische republiek"
          - "Delische plantagegebied"
          - "West-Javaanse Bandoeng"
          - "Joodse gedeelte"
          - "de Europese landen"

      - rule_id: PLC_INC002
        description: Include metonymy references
        conditions:
          - "Metonymy does not disqualify a place name. Metonyms can be double tagged as PLACE as well as ORGANISATION and/or BEING depending on context"
        examples:
          - "Moskou weigert"
          - "Den Haag bepaalt"
          - "Nederland scoort"

      - rule_id: PLC_INC003
        description: Include articles for generic place references
        conditions:
          - "Articles are included when the place reference is generic"
        examples:
          - "{De Laan van Meerdervoort}"
          - "in {het comptoir}"

      - rule_id: PLC_INC004
        description: Tag places embedded in other entities
        conditions:
          - "When a place appears within a BEING description, denomination, or other entity, tag it separately"
        examples:
          - "Maria uit Bengaal → tag 'Bengaal' as PLACE"
          - "Jan van Amsterdam → tag 'Amsterdam' as PLACE"
          - "de Koning van Pruisen → tag 'Pruisen' as PLACE"

    exclusion_rules:
      - rule_id: PLC_EXC001
        description: Do not include articles for specific place names
        rationale: Articles are not part of proper place names and often indicate that the place refer to an OBJECT, ORGANISATION or other entity
        examples:
          - "Amsterdam (not 'de Amsterdam')"

      - rule_id: PLC_EXC002
        description: Do not tag representatives of places as places
        rationale: These are being or organization references
        examples:
          - "De ambassadeur van Nederland (De ambassadeur = BEING, Nederland = PLACE separately)"

  # 3.3 ORGANISATION
  - entity_type: ORGANISATION
    key: ORG
    label: Organisation
    description: >-
      Organizations including companies, institutions, governments, branches,
      associations, legislative bodies, political parties, military forces,
      sports teams, meetings, bands, religious orders, and ships.

    subcategories:
      - name: companies
        key: COMP
        label: Companies
        description: Studio, Bank, etc.
        examples: 
          - "Philips"
          - "ING"
          - "Shell"

      - name: branches
        key: BRANCH
        label: Branches
        description: Departments and organizational branches
        examples:
          - "ING Rotterdam"
          - "Rekenkamer Gemeente Rotterdam"
          - "Afdelingsbestuur NVM afdeling Arnhem"
          - "NVM Arnhem"

      - name: associations
        key: ASSOC
        label: Associations
        description: Coöperatie, Markt (if not place)
        examples:
          - "NVM"
          - "de vakbond"

      - name: public_facilities
        key: PUBFAC
        label: Public Facilities
        description: School, Universiteit (if not already tagged as place)
        examples:
          - "Middelbare school"
          - "Technische Universiteit Delft"

      - name: legislative_body
        key: LEGI
        label: Legislative Bodies
        description: Parliamentary bodies, legislatures, councils, chambers, etc.
        examples:
          - "Tweede Kamer"
          - "Staten-Generaal"

      - name: grand_residence
        key: GRANDRES
        label: Grand Residences
        description: Paleis (when referring to institution)
        examples:
          - "Paleis op de Dam"  # ADDED: when referring to institution, otherwise tagged as PLACE

      - name: printer
        key: PRINTER
        label: Printers
        description: Printing houses
        examples:
          - "Drukkerij De Standaard"
          - "Uitgeverij Van Dale"

      - name: news_agency
        key: NEWSAG
        label: News Agencies
        description: News agencies and media organizations when referred to as an entity that takes actions or is a place of employment rather then one being consumed (the latter is tagged as TEXTUAL_REFERENCE)
        examples:
          - "De NRC besloot"
          - "medewerkers bij De Expres"
          - "(cor. Volkskrant)"
          - "hoofdredacteur van het Bataviaasch Nieuwsblad"

      - name: media_campaign
        key: MEDIACAMP
        label: Media Campaigns
        description: PR campaigns
        examples:
          - "Hij organiseert {de PR campagne}"

      - name: factory
        key: FACTORY
        label: Factories
        description: Manufacturing facilities as organizations
        examples:
          - "de fabriek van Philips"
          - "de Shell raffinaderij"

      - name: political_party
        key: POLPART
        label: Political Parties  
        description: Political parties
        examples:
          - "CDA"
          - "VVD"
          - "PvdA"
          - "Partij Nasional Indonesia"

      - name: international_organisation
        key: INTORG
        label: International Organizations
        description: International organizations
        examples:
          - "Verenigde Naties"
          - "Europese Unie"

      - name: resistance_movement
        key: RESIST
        label: Resistance Movements
        description: Resistance movements
        examples:
          - "Het verzet"
          - "Kelompok Gerilya"

      - name: authorities
        key: AUTH
        label: Authorities
        description: >-
          Government, Ministries, Councils, Courts
        examples:
          - "Ministerie van Financiën"
          - "Overzeese Rijksdelen"
          - "Buitenlandse Zaken"
          - "Raad voor Aangelegenheden van Indonesië (RAVI)"

      - name: dynasties
        key: DYN
        label: Dynasties
        description: Royal and ruling dynasties
        examples:
          - "Omajjaden dynastie"
          - "Nasriden dynastie"
          - "Hashemieten dynastie"
          - "Oranje-Nassau"

      - name: military_forces
        key: MILFOR
        label: Military Forces
        description: Army, Army units
        examples:
          - "Indische leger"
          - "KNIL"
          - "Mariniers"
          - "Luchtmacht"
          - "Marine"
          - "Infanterie"
          - "Korps Speciale Troepen"
          - "Tijdelijke Wet {Koninklijke Landmacht}" # note: the entire string is also tagged as TEXTUAL_REFERENCE

      - name: sports_team
        key: SPORTTEAM
        label: Sports Teams
        description: Sports teams
        examples:
          - "Ajax"
          - "Feyenoord"
          - "PSV"

      - name: sport_tournament
        key: SPORTTOUR
        label: Sports Tournaments 
        description: Championship, Match
        examples:
          - "EK 2020"
          - "WK 2018"
          - "Olympische Spelen 2024"

      - name: meeting
        key: MEET
        label: Meetings
        description: Conference
        examples:
          - "VN Klimaatconferentie"
          - "G20 Top"
          - "de OM vergadering"

      - name: band_orchestra
        key: BANDORCH
        label: Bands and Orchestras
        description: Musical groups
        examples:
          - "Dé Carels"
          - "Het Metropole Orkest"

      - name: theatre_group
        key: THEATREG
        label: Theatre Groups
        description: Theatre groups
        examples:
          - "Toneelgroep Amsterdam"
          - "Het Nationale Toneel"

      - name: religious_order
        key: RELIGORD
        label: Religious Orders
        description: Religious orders
        examples:
          - "Jezuïeten"
          - "Franciscanen"
          - "Dominicanen"
          - "Shattariyah"
          - "Sufi orde"
          - "Katholieke Kerk"

      - name: ship
        key: SHIP
        label: Ships
        description: Uniquely named ships or vehicles (not model names! those are THING)
        examples:
          - "Stoomschip Sumatra"
          - "Hr.Ms. De Ruyter"


    inclusion_rules:
      - rule_id: ORG_INC001
        description: Tag branches with placenames without prepositions
        conditions:
          - "Placenames indicating branches are only tagged if there is no preposition between the organization and the placename"
          - "also tag the place separately"
        examples:
          - "ING Rotterdam → ORGANISATION (full), Rotterdam → PLACE"
          - "Rekenkamer Gemeente Rotterdam → ORGANISATION (full), Rotterdam → PLACE"
          - "NVM Arnhem → ORGANISATION (full), Arnhem → PLACE"

      - rule_id: ORG_INC002
        description: Tag frequently repeated references to organisational groups
        conditions:
          - "Frequently repeated references to denominations which refer to organisations are tagged"
        examples:
          - "aen generael en raden (raden = council itself)"

      - rule_id: ORG_INC003
        description: tag the place as part of the organisation when the organisation is defined by its location. The place is also tagged separately WITH double tagging
        examples:
          - "Rechtbank te Carcassonne → ORGANISATION (Rechtbank te Carcassonne), PLACE (Carcassonne)"
          - "Regering van Oostenrijk → ORGANISATION (Regering van Oostenrijk), PLACE (Oostenrijk)"

    exclusion_rules:
      - rule_id: ORG_EXC001
        description: Do not include articles
        rationale: Articles are not part of organization names
        examples:
          - "Tweede Kamer (not 'de Tweede Kamer' in annotation)"

      - rule_id: ORG_EXC002
        description: Do not tag abbreviations separately
        rationale: Abbreviations are tagged with full name or separately as needed
        examples:
          - "{Nederlandse Vereniging van Makelaars (NVM)}" # NVM is not tagged separately from Nederlandse Vereniging van Makelaars"
          - "{Koninklijk Nederlands Indisch Leger (KNIL)}" # KNIL is not tagged separately from Koninklijk Nederlands Indisch Leger

      - rule_id: ORG_EXC003
        description: Do not tag groups lacking formal structure or that are referred to in a generic sense
        rationale: These are considered denominations
        examples:
          - "De jongerenbeweging"
          - "De vakbond bewegingen"

      - rule_id: ORG_EXC004
        description: Do not tag representatives as organisations
        rationale: These are beings or denominations
        examples:
          - "De Minster van Buitenlandse Zaken (De Minister van Buitenlandse Zaken = BEING, Buitenlandse Zaken = organisation separately)"

      - rule_id: ORG_EXC005
        description: Tag publications as textual references not organisations
        conditions:
          - "When the publication itself is referenced, not the company"
        examples:
          - "In de NRC stond ... (NRC = textual reference)"

      - rule_id: ORG_EXC006
        description: Do not tag general model names of ships/vehicles or other things as organisations
        rationale: These are considered THING
        examples:
          - "Boeing 747 (not 'Boeing 747-400')"
          - "Tesla Model S"

  # 3.4 DENOMINATION
  - entity_type: DENOMINATION
    key: DEN
    label: Denomination
    description: >-
      Ethnicity, profession, religion, demonym, ideology, language, community
      references. Includes adjectives referring to places/organisations/religions,
      demonyms, languages, pejorative terms, professions, ideological affiliations,
      and group references. With double tagging, denominations within person names (BEING)
      or other entities are also tagged separately.

    subcategories:
      - name: adjective_phrases
        key: ADJPHR
        label: Adjective Phrases
        description: Phrases containing adjectives referring to place/organisation/religion/ideology/language/community
        examples:
          - "Islamitische gemeenschap"
          - "Marxistische overtuiging"
          - "Turkse taal"

      - name: demonym
        key: DEMO
        label: Demonym
        description: References to people from a place
        examples:
          - "Chinees"
          - "China"
          - "Arabier"
          - "Westerlingen"
          - "Bandunger"
          - "Brabander"
          - "Europeanen"
          - "Europeesche dame"
          - "Indonesische jeugd"

      - name: titles_ranks
        key: TITLERANK
        label: Titles and Ranks
        description: >-
          Military, noble, and professional titles (now tagged separately from person names)
        examples:
          - "Kapitein"  # ADDED: Tag separately from person name (BEING)
          - "Gouverneur-Generaal"
          - "President"
          - "Sultan"
          - "Jonker/Jonkheer"  # ADDED: Noble title
          - "Lieutenant"
          - "Resident"
          - "Vice admiraal"

      - name: language
        key: LANG
        label: Languages
        description: Language names as nouns
        examples:
          - "Duits"
          - "Fries"
          - "Azerbeidzjaans"

      - name: abstract_bureaucratic
        key: ABBUREAU
        label: Abstract Bureaucratic Terms
        description: >-
          Implicitly refer to ideological groups or departments, typical for
          late 20th/early 21st century texts
        examples:
          - "mix van communicatie- en beleidsverantwoordelijken"

      - name: time_zone
        key: TIMEZONE
        label: Time Zones
        description: Time zone references
        examples:
          - "Zuid-Sumatratijd"

      - name: religion
        key: RELIGION
        label: Religions
        description: Religion names as nouns
        examples:
          - "Protestantisme"

      - name: pejorative
        key: PEJOR
        label: Pejorative Terms
        description: Pejorative terms
        examples:
          - "Slaaf"
          - "Koelie"
          - "Inlander"
          - "Zwarte"
          - "Boschnegers"
          - "Roodharige barbaren"
          - "Totok"
          - "Mohammedanen"

      - name: profession
        key: PROF
        label: Professions
        description: Professional titles and roles
        examples:
          - "Schrijver"
          - "Klerk"
          - "Agent"
          - "Ministers"
          - "Vertegenwoordiger"
          - "Hoofden"
          - "Mandoer"

      - name: religious_ideological_members
        key: RELIDEO
        label: Religious and Ideological Members
        description: Members of religions/ideologies
        examples:
          - "Kapitalist"
          - "Communist"
          - "Moslim"
          - "Christen"
          - "Pro-Russisch"
          - "Anti-abortus"

      - name: group_references
        key: GROUPREF
        label: Group References
        description: General nouns referring to groups through prepositions/possessives
        examples:
          - "{Volk van West-Irian} (in addition: West-Irian = place)"
          - "De {heren van de Volkrant} (in addition: Volkrant = organisation)"
          - "{KPN medewerkers} (in addition: KPN = organisation)"
          - "{die van sammadang} (in addition: sammadang = place)"
          - "{Baccherachs volck} (in addition: Baccherach = BEING)"

    inclusion_rules:
      - rule_id: DEN_INC001
        description: Tag both adjective and noun in denomination phrases
        conditions:
          - "When a phrase contains an adjective referring to place/organisation/religion/ideology"
        examples:
          - "Islamitische gemeenschap → DENOMINATION (full)"
          - "Marxistische overtuiging → DENOMINATION (full)"
          - "Indonesische jeugd → DENOMINATION (full)"

      - rule_id: DEN_INC002
        description: Tag profession/pejorative when appears alone without name
        conditions:
          - "Reference to profession or pejorative occurs without a personal name"
          - "Follow rules from section 3.1.A"
          - "double tag the denomination separately in case this also concerns a specific reference to a person"
        examples:
          - "De minister ('De minister' = BEING, 'minister' = denomination)"
          - "De slaaf rent weg ('de slaaf' = BEING, 'slaaf' = denomination)"

      - rule_id: DEN_INC003
        description: Tag titles and denominations within person names separately
        conditions:
          - "When a person name includes a title, rank, or denomination, tag the full person AND the title/denomination separately"
        examples:
          - "Kapitein Jonker → BEING (full), Kapitein → DENOMINATION"
          - "Gouverneur-Generaal van Starkenborgh → BEING (full), Gouverneur-Generaal → DENOMINATION"
          - "de Javaanschen Cap:n Soeta Wangsa → BEING (full or name), Javaanschen Cap:n → DENOMINATION"

      - rule_id: DEN_INC004
        description: tag denominations in organisation names
        rationale: Organisations have formal structure
        examples:
          - "Nederlandse groep --> organisation, 'Nederlandse' is denomination" 
          - "De Nederlandse Bank --> organisation, 'Nederlandse' is denomination"

      - rule_id: DEN_INC005
        description: Tag group references with denominations
        rationale: These are denominations referring to groups
        examples:
          - "Volk van West-Irian → DENOMINATION (Volk van West-Irian), PLACE (West-Irian)"
          - "KPN medewerkers → DENOMINATION (KPN medewerkers), ORGANISATION (KPN)"

      - rule_id: DEN_INC006
        description: Tag currencies as denominations in case denominations are used
        rationale: Currencies are also textual references or quantities or objects 
        examples:
          - "Spaenschen reael → DENOMINATION (Spaenschen), TEXTUAL REFERENCE (Spaenschen reael)"

      - rule_id: DEN_INC007
        description: Tag denominations with numerals as quantities, but also tag denomination
        rationale: These are quantities, but the denomination part is also tagged separately
        examples:
          - "Twee Nederlanders → QUANTITY (full), Nederlanders → DENOMINATION"
          - "Een twintigtal soldaten → QUANTITY (full), soldaten → DENOMINATION"


  # 3.5 QUANTITY
  - entity_type: QUANTITY
    key: QTY
    label: Quantity
    description: >-
      Quantities including currency, merchandise counts, people counts,
      age, school class, weapons, settlements, area, distance, calibre,
      enumerations, carat, degree, and weight. With double tagging,
      objects within quantity expressions are also tagged as THING.

    subcategories:
      - name: currency
        key: CURRQ
        label: Currency
        description: Monetary amounts
        examples:
          - "ƒ 1.50"
          - "twee gulden" # Also tag 'gulden' as THING

      - name: merchandise
        key: MERCHQ
        label: Merchandise
        description: Counts of goods
        examples:
          - "een lepel"  # Also tag 'lepel' as THING
          - "twee kandelaars"  # Also tag 'kandelaars' as THING
          - "drie schootels"  # Also tag 'schootels' as THING

      - name: people
        key: PPLQ
        label: People
        description: Counts of people
        examples:
          - "23: inlandsche zieken"  # ALSO tag 'inlandsche zieken' as DENOMINATION

      - name: troops
        key: TROOPQ
        label: Troops
        description: Military unit quantities
        examples:
          - "vijfhonderd weerbare mannen" # ALSO tag 'weerbare mannen' as DENOMINATION
          - "4. van de macassaeren" # ALSO tag 'macassaeren' as DENOMINATION

      - name: age
        key: AGEQ
        label: Age
        description: Age expressions
        examples:
          - "honderdjarige leeftijd"

      - name: school_class
        key: SCHCLASSQ
        label: School Class
        description: School class levels
        examples:
          - "derde klas middelbare school" # ALSO tag 'middelbare school' as ORGANISATION

      - name: weapon
        key: WEAPQ
        label: Weapon
        description: Weapon quantities
        examples:
          - "ruijm 1000. stx:s schiet geweeren" # ALSO tag 'schiet geweeren' as THING

      - name: settlement
        key: SETTQ
        label: Settlement
        description: Settlement counts
        examples:
          - "twee negorijen" # ALSO tag 'negorijen' as PLACE

      - name: area
        key: AREAQ
        label: Area
        description: Area measurements
        examples:
          - "een landt van 40 vierkante mijlen"

      - name: distance
        key: DISTQ
        label: Distance
        description: Distance measurements
        examples:
          - "25 mijlen"
          - "2 mijl in zee"
          - "drie â vier dagen varens"
          - "5 uuren oostwaartsheenen"

      - name: calibre
        key: CALQ
        label: Calibre
        description: Weapon calibre
        examples:
          - "kaliber 6.5"

      - name: enumeration
        key: ENUMQ
        label: Enumeration
        description: Lists of counted items
        examples:
          - "3 schootels, zadel, 2 stijgh beugels" # ALSO tag 'schootels', 'zadel', 'stijgh beugels' as THING

      - name: carat
        key: CARATQ
        label: Carat
        description: Gold/gem quality measure
        examples:
          - "gouden sieraden 22, 23 en 24 Kt" # ALSO tag 'gouden sieraden' as THING

      - name: degree
        key: DEGQ
        label: Degree
        description: Degree measurements
        examples:
          - "10 graed noorder breedte"

      - name: weight
        key: WEIGHTQ
        label: Weight
        description: Weight measurements
        examples:
          - "14940 石 quiksilver" # ALSO tag 'quiksilver' as THING

    inclusion_rules:
      - rule_id: QTY_INC001
        description: Infer single items in enumerations
        conditions:
          - "In enumerations, items without explicit numbers are assumed to be singular"
        examples:
            - "3 schootels, zadel, 2 stijgh beugels (zadel = 1 saddle, tagged as quantity)" # ALSO tag 'schootels', 'zadel', 'stijgh beugels' as THING
      - rule_id: QTY_INC002
        description: Tag denominations with numerals as quantities
        conditions:
          - "Denominations preceded by numerals or quantitative adjectives become quantities"
          - "ALSO tag the denomination separately"
        examples:
          - "Twee Nederlanders → QUANTITY (full), Nederlanders → DENOMINATION"
          - "Een twintigtal soldaten → QUANTITY (full), soldaten → DENOMINATION"

      - rule_id: QTY_INC003
        description: Tag travel time as distance not temporal
        conditions:
          - "Time expressions measuring distance are quantities not temporal references"
        examples:
          - "{drie â vier dagen varens} (not temporal reference)"
          - "{5 uuren oostwaartsheenen} (not temporal reference)"

      - rule_id: QTY_INC004
        description: Tag objects within quantity expressions as THING
        conditions:
          - "When a quantity expression includes a countable object, tag the full quantity AND the object separately as THING"
        examples:
          - "een lepel → QUANTITY (full), lepel → THING"
          - "twee kandelaars → QUANTITY (full), kandelaars → THING"
          - "3 schootels → QUANTITY (full), schootels → THING"
          - "zadel → QUANTITY (inferred 1), zadel → THING"
      - rule_id: QTY_INC005
        description:  Tag textual references as quantities if they contain numbers
        rationale: Written sources are textual references but can also contain quantities
        examples:
          - "2 brieven --> textual reference (full), quantity (full)"
      - rule_id: QTY_INC006
        description: Associated organisations are tagged as quantities too when appearing within quantity expressions
        rationale: Organisation is tagged separately WITH the quantity
        examples:
          - "derde klas middelbare school → QUANTITY (full), ORGANISATION (middelbare school)"

  # 3.6 TEMPORAL_REFERENCE
  - entity_type: TEMPORAL_REFERENCE
    key: TMP
    label: Temporal Reference
    description: >-
      Temporal references including days, dates, campaigns/wars, holidays,
      canonised periods, genitives, and temporal adjectives.

    subcategories:
      - name: days
        key: DAY
        label: Days
        description: References to specific days
        examples:
          - "Deze Maandag"

      - name: days_of_week
        key: DOW
        label: Days of the Week
        description: Weekday names
        examples:
          - "Maandag"

      - name: deictic_temporal
        key: DEIC
        label: Deictic Temporal Pronouns
        description: Deictic temporal pronouns
        examples:
          - "gisteren"
          - "morgen"
          - "vandaag"
          - "overmorgen"
          - "eergisteren"
          - "Hedenmiddag"
          - "nu"
          - "gisteravond"

      - name: dates
        key: DATE
        label: Dates
        description: Dates in every calendar
        examples:
          - "Vrijdag 8 November 1957"

      - name: campaigns_wars
        key: CAMPWAR
        label: Campaigns and Wars
        description: Campaigns/wars when referring to time periods
        examples:
          - "Hongitochten"
          - "Twee Wereldoorlog"
          - "1ste Nederlandse militaire actie"

      - name: holidays_festivals
        key: HOLFEST
        label: Holidays and Festivals
        description: Holiday and festival names
        examples:
          - "Nieuwjaar"
          - "geboortedag van de Profeet Mohammad"
          - "Heldendag"

      - name: canonised_periods
        key: CANPER
        label: Canonised Historical Periods
        description: Canonised historical periods (any historiography)
        examples:
          - "Middeleeuwen (Western Europe)"
          - "Periode van de Strijdende Staten (China)"
          - "Zaman Hindu-budis (Indonesia)"

      - name: genitives
        key: GENIT
        label: Genitives
        description: Genitive temporal expressions
        examples:
          - "9 dezer"
          - "Dezer dagen"

      - name: temporal_adjectives
        key: TEMPADJ
        label: Temporal Adjectives
        description: Temporal adjectives before place names
        examples:
          - "eighteenth-century Europe"

    inclusion_rules:
      - rule_id: TMP_INC001
        description: Always tag days
        conditions:
          - "Days are always tagged as temporal references, even when part of a full date"
        examples:
          - "Afgelopen Vrijdag (tagged)"
          - "Vrijdag 8 November 1957 (entire date tagged, including Vrijdag)"

      - rule_id: TMP_INC002
        description: Tag campaigns/wars as temporal when referring to period
        conditions:
          - "Campaign/War is tagged as temporal reference when contextualised to refer to time period"
        examples:
          - "Hongitochten"
          - "Tweede Wereldoorlog"

  # 3.7 TEXTUAL_REFERENCE
  - entity_type: TEXTUAL_REFERENCE
    key: TXT
    label: Textual Reference
    description: >-
      References to written sources, documents, laws, titles of cultural works,
      inventory numbers, accounts, currency types, telephone numbers, URLs,
      programs, policies, agreements, sanctions, statements, laws, surveys,
      stocks, registers, meeting minutes, activities with recorded minutes,
      slogans, proverbs, flags, and honours.

    subcategories:
      - name: radio_frequencies
        key: RADIO
        label: Radio Frequencies
        description: Radio frequencies
        examples:
          - "FM 105.3"

      - name: board_games
        key: BOARD
        label: Board Games
        description: Board game names
        examples:
          - "Monopoly"

      - name: reports
        key: REPORT
        label: Reports
        description: Reports and documents
        examples:
          - "Jaarverslag 1995"

      - name: cultural_titles
        key: CULTITLE
        label: Cultural Titles
        description: Titles of books, songs, movies, pamphlets, records, manuscripts, musicals, programs, magazines, newspapers, journals
        examples:
          - "Stabat Mater van Pergolesi --> TEXTUAL Reference (full), BEING (Pergolesi)"

      - name: inventory_numbers
        key: INVNUM
        label: Inventory Numbers
        description: Archive inventory numbers
        examples:
          - "AVS INV. 61855-3"

      - name: accounts
        key: ACCOUNT
        label: Accounts
        description: Bank and postal accounts
        examples:
          - "Giro 158225"
          - "postgirorekening No. 400"

      - name: currency_types
        key: CURRTYPE
        label: Currency Types
        description: Currency types (not amounts)
        examples:
          - "Spaenschen reael --> TEXTUAL REFERENCE (full), DENOMINATION (Spaenschen)"
          - "reael pedangh ofte rycxdaelder --> TEXTUAL REFERENCE (full), DENOMINATION (pedangh)"

      - name: mint_runs
        key: MINTRUN
        label: Mint Runs
        description: Print/mint runs of currency
        examples:
          - "aan gehaalde staven, namentlijk No 72, 73, 74, 76" # staven is also tagged as THING

      - name: telephone_numbers
        key: TELNUM
        label: Telephone Numbers
        description: Telephone numbers
        examples:
          - "Tel. 020 -72 84 61"

      - name: mailing_lists
        key: MAILLIST
        label: Mailing Lists
        description: Mailing lists and message systems
        examples:
          - "zij organiseerde de {'message box'}"

      - name: academic_references
        key: ACADREF
        label: Academic References
        description: Academic citations
        examples:
          - "(Scheveningen, 1914)"
          - "(Scholte, 1995)"

      - name: page_numbers
        key: PAGENUM
        label: Page Numbers
        description: Page number references
        examples:
          - "Pagina 39"

      - name: religious_texts
        key: RELTEXT
        label: Religious Texts
        description: Religious text references
        examples:
          - "Surah 17:19"

      - name: urls
        key: URL
        label: URLs
        description: Web URLs
        examples:
          - "www.colonialarchitecture.eu"
      - name: email_addresses
        key: EMAILAD
        label: Email Addresses
        description: Email addresses
        examples:
          - "info@colonialarchitecture.eu"

      - name: programs
        key: PROGRAM
        label: Programs
        description: Software or organizational programs
        examples:
          - "Microsoft Word"
          - "Adobe Photoshop"

      - name: policies 
        key: POLICY
        label: Policies
        description: Publicly announced policies
        examples:
          - "passen-stelsel"
          - "pers breidel"
          - "Manokwari-plan" # also tagged as PLACE (Manokwari)
          - "Pax Neerlandica" # also tagged as DENOMINATION (Neerlandica)
          - "non-coöperatie"
          - "presidentieel besluit No. 9" # also tagged as DENOMINATION (presidentieel)
          - "Handvest der Verenigde Naties" # also tagged as ORGANISATION (Verenigde Naties)

      - name: agreements
        key: AGREEMENT
        label: Agreements
        description: Written agreements
        examples:
          - "Linggarjati-overeenkomst" # also tagged as PLACE (Linggarjati)
          - "Renville-overeenkomst" # also tagged as PLACE (Renville)

      - name: sanctions
        key: SANCTION
        label: Sanctions
        description: Written sanctions
        examples:
          - "Poenale Sanctie"

      - name: statements
        key: STATEMENT
        label: Statements
        description: Written/recorded statements
        examples:
          - "communicatie-uitingen"

      - name: laws
        key: LAW
        label: Laws
        description: Legal references
        examples:
          - "artikel 156 alinea 2"

      - name: land_surveys
        key: LANDSURVEY
        label: Land Surveys
        description: Land survey documentation
        examples:
          - "verponding No. 63"

      - name: stocks
        key: STOCK
        label: Stocks
        description: Stock certificates and registered numbers
        examples:
          - "Controleursgrant no. 3"
          - "acte ddo. 12 December 1927 No. 59" # also tagged as DATE (12 December 1927)

      - name: advertisement_registers
        key: ADVREGISTER
        label: Advertisement Registers
        description: Registers of advertisements
        examples:
          - "No. 245 44 regels" # also tagged as QUANTITY (44 regels)

      - name: meeting_minutes
        key: MEETINGMINUTES
        label: Meeting Minutes
        description: Meeting minutes
        examples:
          - "Notulen van de vergadering" # also tagged as ORGANISATION (de vergadering)

      - name: recorded_activities
        key: RECORDACT
        label: Recorded Activities
        description: Activities with recorded minutes/reports if not referred to as an event (TEMPORAL_REFERENCE) or organisation (ORGANISATION)
        examples:
          - "Conference"
          - "Forum"
          - "Concert"

      - name: slogans
        key: SLOGAN
        label: Slogans
        description: Political or advertising slogans
        examples:
          - "Eenheid in verscheidenheid"
          - "Just do it"
          - "Het kan wèl!"

      - name: proverbs
        key: PROVERB
        label: Proverbs
        description: Proverbs and adages (not idioms)
        examples:
          - "Wie honing wil eten moet lijden dat de bijen hem steken"
          - "j'en passe et des meilleurs"

      - name: flags and banners
        key: FLAG
        label: Flags and Banners
        description: Flag and banners
        examples:
          - "Rood, wit, blauw" # also tagged as THING
          - "Bendera Kokki" # also tagged as THING

      - name: honours
        key: HONOUR
        label: Honours and Awards
        description: Titles of honours and awards
        examples:
          - "Ridder in de orde van Oranje Nassau" # also tagged as ORGANISATION (Oranje Nassau) and DENOMINATION (Ridder)

    inclusion_rules:
      - rule_id: TXT_INC001
        description: Tag currency types as textual references
        conditions:
          - "Currency types (not amounts) are textual references"
        examples:
          - "Spaenschen reael" # ALSO tag 'Spaenschen' as DENOMINATION

      - rule_id: TXT_INC002
        description: Tag publications as textual references not organisations
        conditions:
          - "When the publication itself is referred to, not the company"
        examples:
          - "In de NRC stond ... (NRC = textual reference)"

      - rule_id: TXT_INC003
        description: Tag recorded activities as textual references, not events or organisations
        conditions:
          - "Activities of which minutes or reports have been recorded"
        examples:
          - "Conference"
          - "Forum"
          - "Concert"

      - rule_id: TXT_INC004
        description: quantities can contain textual references
        rationale: Written sources are textual references but can also contain quantities
        examples:
          - "2 brieven --> textual reference (full), quantity (full)"

    exclusion_rules:
      - rule_id: TXT_EXC001
        description: Do not tag currency amounts as textual references
        rationale: Amounts are quantities
        examples:
          - "ƒ 1.50 (= quantity)"

      - rule_id: TXT_EXC002
        description: Do not tag news agencies as textual references when they are referred to as organisations
        rationale: Context determines if organisation or textual reference
        examples:
          - "De NRC besloot ... (NRC = organisation)"
          - "In de NRC stond ... (NRC = textual reference)"

  # 3.8 THING
  - entity_type: THING
    key: THG
    label: Thing
    description: >-
      Physical objects, goods, items, and countable entities. This category
      captures concrete objects within quantity expressions, enumerations,
      and descriptions. THING entities typically appear as double-tagged
      elements within QUANTITY annotations.

    subcategories:
      - name: household_items
        key: HOUSEHOLD
        label: Household Items
        description: Domestic objects and utensils
        examples:
          - "lepel"  # spoon
          - "schootels"  # plates
          - "kandelaars"  # candlesticks
          - "zadel"  # saddle
          - "stijgh beugels"  # stirrups

      - name: furniture
        key: FURNIT
        label: Furniture 
        description: Furniture and fixtures
        examples:
          - "tafel"  # table
          - "stoel"  # chair
          - "kast"  # cabinet

      - name: tools_equipment
        key: TOOLS
        label: Tools and Equipment
        description: Tools, instruments, and equipment
        examples:
          - "hamer"  # hammer
          - "zaag"  # saw
          - "mes"  # knife

      - name: clothing_textiles
        key: CLOTH
        label: Clothing and Textiles
        description: Clothing and fabric items
        examples:
          - "hemd"  # shirt
          - "broek"  # trousers
          - "doek"  # cloth

      - name: containers
        key: CONTAINER
        label: Containers and Vessels
        description: Containers and vessels
        examples:
          - "kist"  # chest
          - "vat"  # barrel
          - "zak"  # bag

      - name: weapons
        key: WEAPON
        label: Weapons
        description: Weapons as physical objects (not quantities of weapons)
        examples:
          - "geweeren"  # rifles
          - "zwaard"  # sword
          - "kanon"  # cannon

      - name: Commodities
        key: COMMOD
        label: Commodities
        description: Trade goods and commodities
        examples:
          - "koffie"  # coffee
          - "suiker"  # sugar
          - "specerijen"  # spices

      - name: animals
        key: ANIMAL
        label: Animals
        description: Animals refered to generically (not with names)
        examples:
          - "paard"  # horse
          - "koe"  # cow
          - "kip"  # chicken

      - name: vehicles
        key: VEHICLE
        label: Vehicles
        description: Transportation vehicles
        examples:
          - "wagen"  # wagon
          - "boot"  # boat
          - "fiets"  # bicycle

      - name: documents_as_objects
        key: DOCOBJ
        label: Documents as Physical Objects
        description: Physical documents (not as textual references)
        examples:
          - "brief"  # letter (when referring to physical object in enumeration)
          - "boek"  # book (when referring to physical object)

    inclusion_rules:
      - rule_id: THG_INC001
        description: Tag objects within quantity expressions
        conditions:
          - "When an object appears in a quantity expression (e.g., 'een lepel'), tag the full quantity AND the object as THING"
        examples:
          - "een lepel → QUANTITY (full), lepel → THING"
          - "twee kandelaars → QUANTITY (full), kandelaars → THING"
          - "3 schootels → QUANTITY (full), schootels → THING"

      - rule_id: THG_INC002
        description: Tag objects in enumerations
        conditions:
          - "Objects listed in enumerations are tagged as THING even without explicit quantities"
        examples:
          - "3 schootels, zadel, 2 stijgh beugels → 'zadel' is THING (and QUANTITY with inferred 1)"

      - rule_id: THG_INC003
        description: Tag countable nouns referring to physical objects
        conditions:
          - "Physical, tangible objects that can be counted or enumerated"
        examples:
          - "de lepel ligt op tafel → 'lepel' is THING (contextual)"
          - "hij droeg een zwaard → 'zwaard' is THING"

    exclusion_rules:
      - rule_id: THG_EXC001
        description: Do not tag abstract concepts
        rationale: THING is for concrete, physical objects only
        examples:
          - "een idee (abstract, not THING)"
          - "liefde (abstract, not THING)"
          - "vrijheid (abstract, not THING)"

      - rule_id: THG_EXC002
        description: Do not tag places or buildings as THING is they are unique
        rationale: These are PLACE entities
        examples:
          - "het huis (PLACE or building context)"
          - "de fabriek (ORGANISATION or PLACE)"

      - rule_id: THG_EXC003
        description: Do not tag people or animals with names as THING
        rationale: These are BEING entities
        examples:
          - "de slaaf Louis (BEING, not THING)"
          - "Lassie (BEING - named animal)"

      - rule_id: THG_EXC004
        description: Do not tag documents when referring to content
        rationale: These are TEXTUAL_REFERENCE when content is meant
        examples:
          - "In de brief stond ... (TEXTUAL_REFERENCE, not THING)"
          - "Het boek vertelt ... (TEXTUAL_REFERENCE, not THING)"
          - "But: '2 brieven op tafel' → 'brieven' can be THING if referring to physical objects"

usage_guidelines:
  - "Always start with the broadest entity (e.g., full person name)"
  - "Then identify and tag constituent semantic parts (title, place, object)"
  - "Use span coordinates to track exact character positions of overlapping entities"
  - "In ambiguous cases, prefer more semantic tags over fewer (enrich the data)"
  - "Document rationale for double-tagging decisions in annotation metadata"
  - "Maintain consistency within a single document or collection"