Namescape Named Entity Annotation

Tagging of named entities in the Namescape project

 

As pointed out in f.i. [van Dalen-Oskam 2012], standard NE tagging is not adequate for a systematic study of the onymic landscape. We define an annotation scheme for named entity information as an extension to the TEI p5 annotation guidelines.

TEI extension for named entities

 

The main reasons to extend TEI instead of using the existing tagging guidelines for names are:
1. We need all searchable properties of names to be ‘inline’, not standoff. This entails the introduction of several additional attributes for tagged named entity occurrences.
2. We prefer one single tag for named entities, one (but different) tag for entity parts, whereas TEI either offers a range of tags (persName, geoName, orgName), or the single tag “name”, which latter choice is inconvenient for querying because tagging name parts would lead to nested “name” tags:



<name type="person">


 <name type="forename">Jan</name>


 <name type="surname'>Janssen</name>


</name>


 


The formal definition of the extension has the form of a TEI ODD file[1] , cf. delivered file “namescape.odd.xml”. We include the documentation from the ODD as a description of the proposed encoding guidelines


TEI modules


header, core, tei, textstructure, analysis, dictionaries, drama, namesdates, figures, iso-fs


Element ne


Namespace: http://www.namescape.nl/


Named Entity




<ns:ne


   xmlns:ns="http://www.namescape.nl/"


   type="person"


   gloss="MAIN CHARACTER"


   structure="forename"


   nymRef="nym7"


   normalizedForm="MICHIEL"


   resolution="plotInternal">


       <ns:nePart type="forename" sex="male">Michiel</ns:nePart>


 </ns:ne>




Belongs to element classes:


model.nameLike


Content model


 


<rng:ref name=”macro.phraseSeq” ></rng:ref>


Attributes


 






Name
Description




Type
Named entity typeData type: text


Possible values:person, location, organisation, misc





nymRef
Reference to the “nym” element in the headerData type: tekst




gloss
For persons, this defines the role of the character; for other entities, it is a subcategorisationData type: tekst




structure
This (redundant) attribute gives the internal structure of a person name, like forename_surname, etcData type: tekst




normalizedForm
Form of the entity without interpunction, genitive “s”, uppercasedData type: tekst




resolution
Does the named entity refer to a plot-internal of plot-external concept?Data type: text


Possible values:plotInternal, plotExternal







 


Element nePart


Namespace: http://www.namescape.nl/


Person name part (forename, surname, addname)


Belongs to element classes:


model.persNamePart


Content model


<rng:ref name=”macro.phraseSeq” ></rng:ref>


Attributes






Name
namespace
description




type

Part type(forename, surname, byname)Data type: text


Possible values:surname, forename, addname





sex

Sex of person referred to; only with forenamesData type: text


Possible values:male, female, unknown





surnameType

“Modern” means established, registered surname according to modern usage; “historical” for surname-like (geonymic, patronymic) designations predating modern practice. “collective” for usage like “the Clintons”, etc.Data type: text


Possible values:modern, historical, collective





normalizedForm

Form of the entity part without interpunction, genitive “s”, uppercasedData type: text






 


Element nym


 


“Nym” is a TEI element. The listNym (in sourceDescription in header) enumerates different entities found in the text – something of a small lexicon of the names in the text. We add a few namescape-specific attributes.




Example entry:


<nym ns:id="nym7" ns:resolution="plotInternal" ns:gloss="MAIN CHARACTER" ns:type="person">


  <usg type="frequency">531</usg>


  <form type="nym">MICHIEL VAN BEUSEKOM</form>


  <form type="witnessed">


    <orth type="original">Michiel</orth>


    <orth type="normalized">MICHIEL</orth>


    <usg type="frequency">501</usg>


  </form>


  <form type="witnessed">


    <orth type="original">Michiels</orth>


    <orth type="normalized">MICHIEL</orth>


    <usg type="frequency">25</usg>


  </form>


  <form type="witnessed">


    <orth type="original">Michiel van Beusekom</orth>


    <orth type="normalized">MICHIEL VAN BEUSEKOM</orth>


    <usg type="frequency">4</usg>


  </form>


  <form type="witnessed">


    <orth type="original">v.B.</orth>


    <orth type="normalized">V.B.</orth>


    <usg type="frequency">1</usg>


  </form>


</nym>




Attributes






name
namespace
description




type
http://www.namescape.nl/
Data type: textPossible values:person, location, organisation, misc




gloss
http://www.namescape.nl/
Data type: text




resolution
http://www.namescape.nl/
Data type: text






Element p


Attributes






name
namespace
description




id
http://www.politicalmashup.nl
Data type:




numTokens
http://www.politicalmashup.nl
Data type:






 


Element docStats


Namespace: http://www.politicalmashup.nl


Belongs to element classes


model.teiHeaderPart


Content model


<rng:zeroOrMore>


<rng:choice>


<rng:ref name=”histogram” ></rng:ref>


<rng:ref name=”parTokensMedian” ></rng:ref>


<rng:ref name=”pagebreaks” ></rng:ref>


</rng:choice>


</rng:zeroOrMore>


 


Element histogram


Namespace: http://www.politicalmashup.nl


Content model


<rng:zeroOrMore>


<rng:element name=”entry” ns=”http://www.politicalmashup.nl” >


<rng:attribute name=”bin” ns=”http://www.politicalmashup.nl” >


<rng:data type=”integer” ></rng:data>


</rng:attribute>


<rng:attribute name=”count” ns=”http://www.politicalmashup.nl” >


<rng:data type=”integer” ></rng:data>


</rng:attribute>


</rng:element>


</rng:zeroOrMore>


Attributes






name
namespace
description




description
http://www.politicalmashup.nl
Data type:






 


Element parTokensMedian


Namespace: http://www.politicalmashup.nl


Content model


<rng:data type=”integer” ></rng:data>


 


Element pagebreaks


Namespace: http://www.politicalmashup.nl


Content model


<rng:data type=”integer” ></rng:data>


 


Element collection


Namespace: http://www.politicalmashup.nl


Belongs to element classes:


model.teiHeaderPart


Content model


<rng:text></rng:text>


 


Element pseudonym-id


Namespace: http://www.politicalmashup.nl


Belongs to element classes:


model.teiHeaderPart


Content model


<rng:text></rng:text>


 


Element cleanParagraphs


Namespace: http://www.politicalmashup.nl


Belongs to element classes:


model.teiHeaderPart


Content model


<rng:text></rng:text>


 


Element genre


Namespace: http://www.politicalmashup.nl


Belongs to element classes:


model.teiHeaderPart


Content model


<rng:text></rng:text>