Relation^3: How to relate text describing relationships with structured encoding of the relationships?

Authors: Galka, Selina Veronika / Vogeler, Georg

Date: Wednesday, 6 September 2023, 9:15am to 10:45am

Location: Main Campus, L 2.202 <campus:measure>


How can personal relationships (“Henry and his brother”) in historical sources be modeled under the aspect of “assertive editing”? The concept of “assertive edition” is based on a multi-layered representation by linking embedded annotations (annotations with TEI/XML) with external information structures (e.g. RDF) in order to model the facts contained in the text as assertions (Vogeler 2019, 315), whereby different possibilities for this linking exist (Boot/Koolen 2021).

For modeling relationships of persons, the TEI proposes the element tei:relation with the attributes @name, @active and @passive. We follow this proposal for the modeling of the relationships themselves using the attributes @type for more precise specification of the relation, e.g. “family”, @name for the type of relation, e.g. “brother” and @mutal, @active or @passive for the persons involved (“#Henry”, “#HenrysBrother”). This modeling approach is simple, can be easily translated to RDF and as much information as possible remains in the TEI/XML data. However, the TEI does not declare a canonical way to link prosopographical information with text. For linking TEI/XML and RDF structures, RDFa can be embedded directly in TEI/XML; tei:xenodata offers a possibility to embed RDF statements in the TEI document. One argument against the use of both of them is the more difficult human readability of the TEI document (Vogeler 2019, 317). As tei:xenodata is defined in the context of the tei:teiHeader, its main purpose is in the storage of metadata and, by this, seems to be out of scope. 

We suggest annotating the text with tei:seg and referring to the @xml:id of the tei:relation with the attribute @ana. @ana seems to be the best solution to refer to statements, as @ref is used to link to an entity (Vogeler 2019, 317) and @ana is a globally available attribute. Schwartz, Gibson and Torabi also dealt with modeling relationships in their projects, where they create a full blown factoid prosopography with TEI. They rely, similar to our approach, on the use of tei:relation to represent the familiar relationships, but the linking method applied is much less granular, as the general factoid model does not require a detailed annotation of the source on which the claim of the factoid is based.

To link the relations and the text, it would be also possible to use tei:annotation - it is following the web-annotation model pointing into the text (annotation/ptr), a referencing method one could call “inward pointing” from the perspective of editorial sciences, while one could call our method, i.e. linking the <relation> to the text as context to which we point via the @ana mechanism from the edited text, “outward pointing”. Both the tei:xenoData and the tei:annotation introduce an additional effort in encoding with not adding more information necessary for our edition: the tei:annotation would add a richer description of responsibility and documentation of the annotation act, while the tei:xenoData approach would add rich linked data semantics which would only help to include external definitions of the relation types, while all other modeling semantics is available in the TEI itself. In the context of our edition this additional effort is not justified. Therefore, we decided to use the most simple encoding pattern. The encoding can be converted via XSLT into both other methods without losing information, see GitHub: MGS. Additionally, the outward pointing method is better aligned with the layered editorial concepts as presented by Elena Pierazzo (2015, 43) or Patrick Sahle (2013:III, 251-340), where the annotation is considered an interpretation of the text passage, i.e. an analytic step based on the text itself. 

With these modeling proposals, no new TEI elements or attributes would need to be defined, all information is available in the TEI/XML and can be mapped to RDF with appropriate transformations.


Boot, Peter and Koolen, Marijn. “Connecting TEI Content Into an Ontology of the Editorial Domain.”. In Graph Data-Models and Semantic Web Technologies in Scholarly Digital Editing, edited by Elena Spadini, Francesca Tomasi, and Georg Vogeler, 9-29. Norderstedt: BoD, 2021.

Pierazzo, Elena. Digital Scholarly Editing: Theories, Models and Methods. Farnham: Ashgate, 2015.

Project: SAWS. Composed of teams working at King’s College London, the Newman Institute, Uppsala, and the University of Vienna. URL: [2023-04-15].

Project: Vanderbilt University, Princeton University, and the Contributor(s), 2014. URL: [2023-04-13].

Project: SPEAR. Syriac Persons Events and Relations. [2023-08-01].

Sahle, Patrick. Digitale Editionsformen. Zum Umgang mit der Überlieferung unter den Bedingungen des Medienwandels. Teil 3: Textbegriffe und Recodierung. Schriften des Instituts für Dokumentologie und Editorik 9. Norderstedt: Books on demand, 2013.

TEI: <relation>. Version 4.6.0. Last updated on 4th April 2023, revision f18deffba. URL: [2023-04-15].

Schwartz, Daniel L., Nathan P. Gibson, und Katayoun Torabi. „Modeling a Born-Digital Factoid Prosopography Using the TEI and Linked Data“. Journal of the Text Encoding Initiative, 21. März 2022.

Vogeler, Georg. “The ‘assertive edition’: On the consequences of digital methods in scholarly editing for historians.” International Journal of Digital Humanities 1 (2019): 309–322.

About the authors

Selina Galka is a research project assistant at the Institute Centre for Information Modelling (University of Graz), currently working in the project Digital Edition of the Memoirs of Countess Schwerin (1684-1732). Her main research interests are digital edition and data modeling.

Georg Vogeler is professor for Digital Humanities at the Institute “Centre for Information Modelling” at the University of Graz. He served in the TEI board of directors in 2018-2019. In the Digital Humanities his research interest lies in Digital Scholarly Editing, Semantic Web technologies, Data Modelling, and application of Data Science to the Humanities.

Contribution Type