Representing Chinese cultural heritage through Digital Humanities tools isn’t straightforward. While there’s been a lot of progress in using Linked Open Data (LOD) to describe artworks, museums, and historical artifacts, Chinese calligraphy—especially Chinese contemporary calligraphy—is still largely missing from the picture. Most existing models are suitable for capturing the contextual elements, for instance who made an artwork, where it was created, or what period it belongs to. However, when it comes to the calligraphy itself, there’s often very little detail. This is especially true since sometimes one cannot even say with certainty that the (brush)strokes and the forms are calligraphy indeed. Hence, it is of the great importance to analyse and represent both the artistic and linguistic layers in order to answer the most delicate and valuable question posed by the WRITE project: Can Chinese contemporary calligraphy still be considered calligraphy in traditional terms, and if yes, to what extent.
Well-known ontologies like CIDOC CRM [1] are powerful tools for representing cultural heritage data, and they’re widely used in museums and cultural heritage projects. However, they can be difficult to reuse in more domain-specific or niche projects, especially when trying to represent artistic forms that are deeply tied to language, gesture, or culture. Similarly, initiatives like the Getty Art & Architecture Thesaurus (AAT) [2] do include concepts for traditional Chinese scripts, art materials and tools, but they don’t go very far when it comes to describing the actual art of calligraphy in its modern or experimental forms.
There are several notable projects that focus on Chinese cultural heritage, but most of them do not produce reusable ontologies or thesauruses for contemporary artistic and calligraphic practices. For instance, CRM-ACA - Ancient Chinese Artifacts Conceptual Reference Model [3] is a CIDOC-CRM-based extension ontology tailored for modeling ancient Chinese artifacts, in particular ancient Chinese ritual bronzes. The Dunhuang Mural Thesaurus [4] is a domain-specific vocabulary developed to classify iconographic elements in the Buddhist murals of the Dunhuang caves, focusing on religious and historical imagery. WuMKG - Wu Men Multimodal Knowledge Graph [5] is a knowledge graph dedicated to the Wu Men School of painting, emphasizing structured information around a specific historical art movement. CICHMKG [6] is a multimodal knowledge graph for Chinese intangible cultural heritage that integrates various data types — text, image, audio — but primarily concentrates on preservation and categorisation rather than offering domain-agnostic semantic models. While all of these are important and valuable contributions, they mainly center on traditional or historical content. None are designed to model the rich, evolving world of contemporary Chinese calligraphy or its intersections with contemporary art.
This is the gap that the ontology and thesaurus developed within the WenDAng project aim to try and bridge, or at least, contribute to filling. The goal is not just to adapt existing standards, but to build on them in a way that supports the nuanced representation of contemporary Chinese calligraphy and related arts. This includes describing not only tangible elements like tools, materials, or types, but also intangible aspects — such as script styles, Chinese cultural concepts and visual elements embedded in the artworks — that are often overlooked in conventional datasets.
In short, this project offers a way to make contemporary Chinese calligraphy and art more visible and accessible in the digital world. Once completed, it will aim to support researchers, artists, and institutions who are working with these forms today, and create a foundation for better representation, discovery, and sharing in the future.
The WRITE ontology is being designed to represent the intangible cultural heritage of Chinese calligraphy as it appears in contemporary artworks, alongside the broader tangible cultural heritage of contemporary arts. The ontology provides a structured way to describe artworks, their contextual elements, and the calligraphy integrated within them. Additionally, it facilitates the identification of similarities and differences between contemporary and traditional calligraphy, enabling a deeper understanding of their evolution and artistic connections.
The WRITE ontology combines established knowledge structures with domain-specific innovations. It reuses classes and properties from Wikidata to ensure interoperability, while also introducing new, calligraphy-focused classes and properties designed to capture the unique aspects of Chinese calligraphy in contemporary art. This hybrid approach enables a rich and nuanced representation of both the artistic practices and cultural significance within the domain.
Here below you will find a detailed, zoomable graph that represents the ontology. This graph was created using Graffoo.
For better understanding and visual clarity, you can explore each class individually down below. Every element is complemented with a short description and a zoomable graph that illustrates how the class in question is connected to other classes in the ontology, through owl:objectProperty, as well as its data and object properties (owl:dataTypeProperty, owl:objectProperty), with respective expected object values.
Visual Arts class [owl:Class] represents individuals that constitute one of the four collections of the WRITE knowledge graph.
Performing Arts class [owl:Class] represents individuals that constitute one of the four collections of the WRITE knowledge graph.
Graffiti class [owl:Class] represents individuals that constitute one of the four collections of the WRITE knowledge graph.
Decorative and Applied Arts class [owl:Class] represents individuals that constitute one of the four collections of the WRITE knowledge graph.
Calli-Writing Unit class [owl:Class] represents a calligraphic, written, pseudo-written, or writing-like component embedded in an artwork. The term has been coined by the domain experts, and it is deliberately broader than calligraphy or writing. The word calligraphy would presuppose the existence of the answer to the WRITE’s research question, while writing would be too broad for this scenario in which the written sections may appear as readable texts, illegible signs, visual forms, gestures, material traces, or conceptual elements.
Series of Creative Worksclass [owl:Class] represents an explicit connection between artworks of the four WRITE collections with the work series they may belong to.
Person class [owl:Class] represents the complex network of agents intended as singles that operate in the context of this project: calligraphers, artists, performers, dancers, musicians, authors, curators, organisers and so on.
Artistic collective class [owl:Class] represents the network of agents intended as groups that operate in the context of this project: crews, dance companies, design studios, architectural firms, tattoo studios, etc.
Organisation class [owl:Class] represents the network of agents intended as institutions that are involved in this project: universities, departments, research centres, museums, art galleries, brand companies etc.
Exhibition class [owl:Class] represents the exhibition(s) in which a certain artwork from the four WRITE collections was displayed.
Literary Work class [owl:Class] represents any literary production connected to a Calli-Writing unit, given that the content of a Calli-Writing unit could have been based on a certain literary piece.
Traditional Artwork class [owl:Class] represents artworks made before 1980 that are somehow related to the contemporary artworks of the four WRITE collections.
Access the ontology
The WRITE ontology is available at this link. The access page is generated using pyLODE.
To access the ontology in RDF (serialized in Turtle), please visit this page or download it here.
The official GitHub repository for the WRITE ontology can be accessed here.
The WRITE Thesaurus is a structured, multilingual vocabulary designed to support the creation, description, and analysis of a knowledge base for contemporary Chinese calligraphy. It plays a critical role in the WenDAng project, where it is used in the data entry process alongside the WRITE ontology, contributing to the generation of a semantically rich RDF dataset. This thesaurus is a controlled vocabulary that reflects the nuanced themes, techniques, and contexts of the practice it documents.
More than just a reference tool, the thesaurus serves as a dynamic framework for knowledge organisation and scholarly inquiry. Its structure is intentionally open-ended and extensible, allowing it to evolve in response to ongoing research and curatorial input. New concepts can be introduced, existing ones refined or restructured, and connections established with external vocabularies as the project’s intellectual scope expands.
Structure
The thesaurus is modelled in RDF using the SKOS (Simple Knowledge Organization System) vocabulary. At its core, the thesaurus comprises:
- A single
skos:ConceptScheme, which represents the top-level vocabulary structure, under which all concepts are organised. - A set of
skos:Collections, used to group related concepts thematically rather than hierarchically. This supports a modular and dynamic approach to organising concepts based on usage context and theme, rather than a rigid hierarchy. - Individual
skos:Concepts, which are the actual terms in the thesaurus.
Each concept [skos:Concept] contains (some or all of) the following fields:
skos:prefLabel: Preferred label for the concept, provided in English@en.skos:altLabel: Alternative labels or synonyms, provided in: English@en, Chinese@zhand transcribed Chinese@zh-Latn.skos:definition: Clear and thorough explanation of the concept.dcterms:source: Citations or sources used to define the concept labels and definitions. Only verified sources are included to ensure academic credibility.
To support rich navigation and knowledge inference, the thesaurus includes:
- Hierarchical relations via
skos:broaderandskos:narrower: Used for defining parent-child relations between concepts within the thesaurus. - Associative relations via
skos:related: Used for defining relations between thematically or semantically connected concepts.
Notably, hierarchies are applied carefully, only where conceptual dependencies are well-defined.
External mapping
To enhance semantic interoperability, where applicable, concepts are linked to terms in other established vocabularies using SKOS mapping properties:
skos:exactMatch: For concepts with full equivalence in external sources.skos:closeMatch: For concepts that are similar but not identical.skos:relatedMatch: For concepts with thematic relation, that are still somehow related.
External vocabularies referenced so far include:
- Getty AAT (Art & Architecture Thesaurus) - especially for Chinese script styles;
- INDIGO Graffiti Thesaurus - for graffiti-specific terminology;
- Chinese Iconography Thesaurus - for visual elements rooted in Chinese cultural tradition.
These mappings extend the thesaurus’s reach and fully follow the FAIR principles.
Thematic groupings
Instead of enforcing a rigid hierarchy, the WRITE thesaurus uses skos:Collection to group concepts based on a theme or a subject. This allows for easier future expansion without restructuring the theusaurus.
The skos:Collections currently available within the thesaurus are:
- Chinese visual elements: 29 terms
- Chinese concepts: 9 terms
- Script styles: 22 terms
- Graffiti types: 7 terms
- Graffiti genres: 6 terms
- Graffiti styles: 18 terms
Each collection uses skos:member to associate relevant concepts.
Design Decisions and SKOS Compliance
The modelling approach used in the WRITE thesaurus intentionally avoids rigid hierarchical classification by relying on skos:Collection and skos:inScheme rather than on skos:hasTopConcept. Therefore:
While skos:hasTopConcept is recommended for creating any SKOS thesaurus, the WRITE thesaurus uses thematic skos:Collection groupings and asserts scheme membership via skos:inScheme. This allows for greater flexibility and accommodates the evolving, non-strictly hierarchical nature of the thesaurus.
Accessibility
The thesaurus has been made accessible through the /DH.arc Vocabularies, created by the Digital Humanities Advanced Research Centre at the University of Bologna. It is available here.
Please find the .ttl file of the first version of the WRITE thesaurus here.
The official GitHub repository for the WRITE thesaurus can be accessed here.
*** The WRITE thesaurus is intended for reuse beyond the immediate project, particularly by communities working on East Asian art, visual culture, or domain-specific digital humanities projects. ***
The WRITE dataset brings together the research data collected by the WRITE team on new forms of contemporary Chinese calligraphy. Organised as a knowledge graph, it connects artworks from the four WRITE collections (Visual Arts, Performing Arts, Graffiti Art, and Decorative and Applied Arts) with their “Calli-Writing” units, people, literary works, exhibitions, series of artworks, and other contextual information. These relationships support the exploration and comparison of the artworks across artistic, linguistic and textual, and socio-political and economic dimensions.
From data entry to knowledge graph
The WRITE domain experts entered their research data into Omeka S using resource templates configured with the ontology's classes and properties. The data were subsequently exported, cleaned, harmonised, and enriched using Python and Jupyter notebooks. This workflow included alignment with the updated WRITE ontology where necessary, normalisation of values and relationships, and links to external authority identifiers. The resulting RDF knowledge graph was validated, serialised in Turtle, and loaded into a triplestore for access through a SPARQL endpoint.
Dataset overview
The latest version of the WRITE dataset includes 24,215 RDF triples, including 315 artworks and 320 “Calli-Writing” units. The dataset is structured using the WRITE ontology, which combines project-specific elements with selected classes and properties from Wikidata, while the WRITE thesaurus provides controlled terminology for culturally specific concepts.
Persistent identifiers
The dataset has the persistent identifier https://w3id.org/write/data.
Individual records are identified using the pattern: https://w3id.org/write/data/item/{id}
SPARQL query example
The following query illustrates one of the dataset's analytical uses: identifying the most frequently recorded script styles in Chinese-language “Calli-Writing” units.
PREFIX ns1: <https://w3id.org/write/ontology#>
PREFIX ns2: <http://www.wikidata.org/prop/>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
SELECT ?scriptStyle (COUNT(DISTINCT ?unit) AS ?count)
WHERE {
?unit a ns1:CalliWritingUnit ;
ns2:P407 <https://www.wikidata.org/entity/Q7850> ;
ns2:P9302 ?style .
?style rdfs:label ?scriptStyle .
}
GROUP BY ?scriptStyle
ORDER BY DESC(?count)
Access and documentation
The WRITE data official GitHub repository provides information about the dataset and its publication.
Its VoID description supplies dataset-level metadata, including information about the URI scheme and vocabularies used.
The knowledge graph can also be explored through the WenDAng – WRITE Digital Archive, a semantic portal developed using Sampo-UI.