In this paper, we explore the workflows and challenges involved in creating ontoterminologies for the study of antiquity, focusing on three distinct humanities use cases developed using the TEDI software. We present a systematic approach to building FAIR (Findable, Accessible, Interoperable, and Reusable) datasets for legal processes in Classical Athens, Archaic Lyric poetry, and the archaeological site of Göbekli Tepe. The paper discusses the workflows we employed to create and define concepts and terms, detailing the conceptual and linguistic challenges encountered, such as aligning ancient and modern terminology and representing uncertainty in historical data. We propose that the iterative nature of TEDI allows for flexible and accurate ontoterminologies that can evolve with scholarly research. Furthermore, we highlight how ontoterminologies can enhance the interoperability of data across diferent humanistic research projects, contributing to the broader vision of the Semantic Web. Our findings emphasize the potential of ontoterminologies to bridge conceptual gaps, clarify domain-specific terms, and enable more efective knowledge exchange within the humanities.
In this short paper, we introduce our methodologies for the creation of ontoterminologies for the
study of antiquity. By presenting our experience building findable, accessible, interoperable, and
reusable (FAIR) [
While “ontology” and “terminology” predate the digital era, both have taken on new meaning within
the proposed Semantic Web [
In order to create humanist ontoterminologies, we use the software TEDI [
In this section, we introduce two ontoterminologies within the domain of Classical studies. The first
of these describes the Classical Athenian legal system by defining the processes and legal bodies [
The first ontoterminology, which describes Classical Athenian legal processes and bodies, is part of
a PhD thesis entitled “Semantic Annotation and Attic Oratory”. This PhD aims to create a semantic
resource for the study of Attic oratory, utilizing a data model built by merging three ontoterminologies
created in TEDI in the standard ontology editor Protégé [
In the next step, Author 1 consulted secondary bibliography for scholarly discussions of the ancient
Greek terms for case types. Through this, she determined the axes of analysis for the concepts designated
by the terms. For example, in the Classical Athenian legal system, diferent magistrates oversee diferent
types of proceedings. The Eponymous Archon ( ) oversees family law, while the Basileus () supervises
suits related to religious infringements [
<Private legal proceeding> + /initiated by submitting a charge in writing/ + /supervised by the Forty/ + /tried by the dikasterion/ + /prosecutor issues oral summons at the Assembly/ + /for causing financial loss/ + /assessable penalty/ + /fine/ + /penalty of one sixth of the sum at issue for failed prosecution/
The final step was to determine the equivalent terms in other languages (English in this case). This step introduces challenges, especially when terms refer to practices which no longer exist in modern legal systems. For instance, the ancient ‘ ’ (literally ‘suit for intentional homicide’) is not a direct equivalent to the US or UK ‘murder in the first degree’. Therefore, Author 1 decided to use more literal English translations of the ancient Greek from scholarship and commentaries, rather than draw false equivalencies between ancient and modern legal systems.
Author 2 followed a similar workflow to model the genre of Archaic Lyric poetry, but experienced
particular challenges due to the fragmentary nature of the genre and debates in current domain research
[
Author 2 began with an examination of the original works of the Archaic Lyric poets. Next, he expanded on his knowledge of the primary texts by reviewing secondary sources, which provided key discussions for the defining characteristics of this poetry. Although this method of textual analysis proved to be challenging due to the multiplicity of perspectives presented by domain experts, it allowed for the identification of axes of analysis by constructing concepts based on discrete combinations of features. These axes of analysis included elements such as content, dialect, performer, and occasion. For example, the “content” axis contained: /with erotic content/, /with gnomic content/, /with imitation content/, /with political content/, and /with war content/. Thus, by prioritizing the primary texts themselves, Author 2 developed a clearer understanding of the sub-genres, ultimately enabling the formulation of an accurate conceptual framework. For example, through this approach, the “erotic elegy” was defined as:
<Archaic Lyric Poem> + /only in ionic dialect/ + /in elegiac couplet/ + /with specific content/ + /with erotic content/
Finally, Author 2 opted to confine the process of designating concepts by terms that have been used
by the bibliography for the literary production of the specific period and the specific poetic genre. To
demonstrate, although these terms appear elsewhere in the ancient Greek literature, e.g. “thre¯nos" and
“hymenaios" [
In comparing these two ontoterminologies, we see how structured, domain-specific approaches can clarify the ancient Greek literary production and its complexity. Despite diferences in subject matter—legal processes and poetic genres—both studies relied on rigorous primary text analysis, secondary scholarship, and systematic classification of essential characteristics. By prioritizing historical accuracy, both authors ensured that their models remain faithful to the original sources while enhancing accessibility for contemporary research and education.
In this section, we introduce our third case study, an ontoterminology within the domain of Prehistoric
Archaeology [
Author 3’s first step was the collection of the archaeological data of the site. She consulted the
archaeological archives of the excavator [
Next, Author 3 defined the concepts in formal language and the terms in natural language. She
searched for the archaeological terms and their definitions in recognised archaeological lexicons [
For example, the term “central T-Shaped pillar” designates a concept with three essential characteristics: that it is a /type of archaeological find/ and not a stratigraphic or relief data; that it is a part of the main architecture of the site (the enclosures); and that it is located in the centre of the enclosure (diferentiation from the lateral pillars of the enclosures). The two first characteristics are inherited from the main concept (<T-Shaped Pillar>) and the last one is a delimiting characteristic. These three together create the formal definition:
<type of archaeological find> + /the main supporting component of the sites architecture/ + /in the centre of the enclosure/.
The choice to develop a unique ontoterminology instead of strictly following standards like CIDOCCRM, while still drawing on their foundational principles, was guided by the structure of the taxonomy and the nature of data from existing research. It was crucial to remain aligned with the intent of previous researchers and present the data accurately and faithfully. In archaeology, precise definition and categorization of artifacts are essential for constructing coherent interpretations. Ontoterminologies created through TEDI ofer the opportunity to insert diefrent date values and characteristics on the objects, something that leads to a more concrete manner of modeling the site and categorizing the data. Through the relationships between the objects of the ontoterminology, we gain a holistic view of the ifndings.
The ultimate goal of the current ontoterminology is to accurately represent archaeological findings without bias or excessive detail, aiming for an objective, logical, and comprehensive view of the research subject. Data was meticulously gathered from publications by site excavators and cross-referenced with various sources to ensure a holistic and detailed representation of the site. This ontoterminology aligns with the eforts of site excavators and researchers, catering to the needs of the archaeological community. In this context, TEDI facilitates the incorporation of visualizations, allowing users to easily export a HTML Onto-Dictionary of terms or proper names which includes images, which is very useful for describing archaeological findings (Figure 3).
In order to demonstrate the eficacy of the three ontoterminologies for antiquity, we provide SPARQL
queries. TEDI allows ontoterminologies to be exported in RDF, which means they can be queried using
the SPARQL Protocol and RDF Query Language (SPARQL) [
One such advantage, showcased through a SPARQL query for the ontoterminology created by Author 1 for Classical Athenian Legal Processes, is the ability to query for essential characteristics. For instance, the following query returns a list of the ancient Greek terms for all legal proceedings supervised by the Thesmothetai, as well as the natural language definitions of these terms (see Figure 4 for results):
?cpt1 otv:diference ?dif1. ?cpt1 skos:definition ?definition. ?dif1 rdfs:label “thesmothetae". ?dif1 rdfs:label ?dif1Name.
FILTER (lang(?termName) = ‘grc’) FILTER (lang(?definition) = ‘en’)}
As previously mentioned, TEDI’s RDF export enables users to query essential characteristics, making it, in the context of ALyrA Ontoterminology, particularly valuable for retrieving the defining features of each type of Archaic Lyric poem. For instance, the following query extracts the essential characteristics that define the concept denoted by the term ‘hymn,’ along with the poetess who composed this type of Archaic Lyric poem (see Figure 5 for results):
PREFIX alyra: <http://www.ontologia.fr/OTB/ALyrA_v.1.0#> PREFIX otv: <http://www.ontologia.fr/OTB/otv#>
FROM <http://ontologia.fr/OTB/ALyrA_v.1.0.rdf> WHERE {?conceptPoem rdfs:label “hymn"@en; otv:diference ?dif. ?dif rdfs:label ?difLabel. ?instancePoem otv:instanceOf ?conceptPoem; alyra:isWrittenBy ?composer. ?composer otv:instanceOf alyra:Archaic_Lyric_Poetess. ?composer rdfs:label ?authorName.}
The third and final ontoterminology, showcasing the archaeological finds at Göbekli Tepe, can be evaluated through the following query, which returns a list of the English terms and their definitions in natural language (see Figure 6 for results):
?concept otv:denotedByTerm ?term. ?term otv:language ?lg. ?term otv:termName ?termName. ?term otv:termDefinition ?definition FILTER (?lg = “en")}
Through our individual experiences creating ontoterminologies for antiquity with TEDI, some mutual challenges emerged. A key challenge was manually determining the relevant terms for each domain. In all three ontoterminologies, NLP methods were determined to not be adequate for term extraction. Although Author 1 and Author 2 worked with a corpus of texts, these texts seldom contain the terms that are used in modern scholarship to describe and categorize the texts. Author 3, working with archaeological findings, experienced a similar challenge of manually determining domain terms by consulting secondary scholarship. All three authors had to undertake the humanistic efort of ensuring our concepts, terms, and definitions aligned with scholarly consensus, which was especially challenging due to the dynamic nature of our fields. The TEDI software proved essential, allowing iterative updates to ontoterminologies and accommodating alternative or regionally specific terms, which enriched our representation of these domains.
Another common challenge was the representation of uncertain attributes, particularly dates. In
the study of antiquity, precise dating is rare, and many objects and concepts exist within broad ranges
or with multiple proposed dates. Representing these ambiguities in a consistent manner was crucial
to ensure we accurately reflected our fields. However, TEDI does not currently have a direct solution
for representing uncertainty (although this is a challenge with standard ontologies as well, leading to
the development of new standards for knowledge graph representation such as RDF-star [
There were, however, some unique challenges faced by each ontoterminology. Author 1 had to
deal with the translation of ancient terms into modern language, even when such terms refer to legal
processes which no longer exist. TEDI’s incorporation of SKOS Core vocabulary made it possible to
include multiple English terms that designated the same concept, and easily denote terms as ‘Preferred’,
‘Alternative’, or ‘Obsolete’ [
Author 2’s domain terms are absent from Archaic Lyric poetry itself but found elsewhere in ancient Greek literature, as forms of Archaic Lyric poetry appear embedded in other ancient literary genres. His conceptual approach, grounded in primary texts and refined through secondary sources, identified key analytical axes such as content, dialect, and occasion, enabling precise classification through unique trait combinations. By restricting terminology to scholarly discussions, he ensured historical accuracy and avoided misapplications across traditions or periods (Figure 9).
Despite the challenges, we believe that our resulting ontoterminologies demonstrate the value of using TEDI for our workflows. Compared to standard ontology editors such as Protégé, TEDI facilitates the creation of complex RDF while accommodating the needs and practices of humanists. Specifically for the creation of terminologies, TEDI streamlines the process of defining concepts through essential characteristics and aligning concepts with terms. The resulting ontoterminologies are queryable for concepts, terms, axes of analysis, and essential characteristics, without requiring users to learn Description Logics. Furthermore, because TEDI allows for the export of ontoterminologies into RDF, it is still possible to view and edit ontoterminologies in Protégé, which was done by Author 1 to merge three ontoterminologies into one database (Figure 10).
TEDI’s flexibility and the iterative nature of our work allowed us to maintain accurate, adaptable ontoterminologies that reflected the latest research while accommodating uncertainty and evolving scholarly perspectives. We also believe that the workflow we each followed for creating ontoterminologies runs in parallel to the methods used in non-digital humanistic research. Once accustomed to TEDI’s interface, humanists can easily incorporate the creation of ontoterminologies into their research, leading to the development of more machine-readable terminologies. Ontoterminologies can be reused to eliminate confusion regarding domain specific terms and to link data across diferent humanistic research projects, allowing us to further fulfill the promise of the Semantic Web.
We gratefully acknowledge the generous funding of the European Union in the context of the “TALOSArtificial Intelligence for the Humanities and the Social Sciences” project under the Horizon Europe Framework Programme (HORIZON-WIDERA-2022-TALENTS-01-01-ERA Chairs); Grant agreement ID: 101087269.
During the preparation of this work, the authors used ChatGPT-4 for grammar and spelling checks. The authors have subsequently reviewed and edited the content and take full responsibility for the publication’s final version.
2005.
URL: