A Feasibility Study With Two Datasets on the Application of the Early Stage HDTO Ontology in Practice

A Feasibility Study With Two Datasets on the Application of the Early Stage HDTO Ontology in Practice Heritage Digital Twins in Practice: A Finno-Ugric Case Study of Music, Garments, and Architecture Antal, Dániel Introduction This document presents a feasibility study for the FinFAIR cascading grant under the ECHOES /European Collaborative Cloud for Cultural Heritage (ECCCH) framework. At the same time, it bridges the findings of OpenMusE on data sharing, governance, and metadata repair with the “digital twin” perspective developed within ECHOES / ECCCH. The study has two goals: first, to summarise the key technical and governance insights gained in OpenMusE; and second, to establish a methodological and semantic bridge toward the next project phase, which will operationalise these methods within the ECCCH by creating digital twins of musical works, performances, and architectural heritage. ĹNote This feasibility study was created in withing the framework of the OpenMusE project, and in the preparation of its exploitation pathway under the Stage 2 Echoes proposal writing. We will keep updating this document as we are extending our dataset and gaining more experience with actual, real life data. Please refer to this document as Daniel Antal (2025): A Feasibility Study With Two Datasets on the Application of the Early Stage HDTO Ontology in Practice DOI: 10.5281/zenodo.17492178 Data Manager: Mester, Anna Márta. Digital twins are conceived as composite digital representations of cultural artefacts or practices—such as a musical work, a performance, or a building—that go beyond a single scanned score, sound recording, or photograph. They integrate multiple perspectives and materials: different arrangements or interpretations of a work, the documents of its first performances, or the architectural plans and subsequent restorations of a building. These digital twins connect tangible and intangible dimensions within one coherent, machineactionable framework. 1 Such complexity requires a carefully structured data curation and rights management workflow. In FinFAIR, we adopt the concept of a data sharing space, following the Open Music Observatory model. Instead of forcing contributors into a single metadata schema or legal framework, the data sharing space allows collaboration on an as-needed and as-permitted basis. Each partner retains control over their own assets while contributing to a common pool of interoperable, ethically governed data. This approach enables the inclusion of a wide variety of materials—free or licensed, public or restricted—without compromising legal compliance or curatorial integrity. Our existing infrastructures, such as the Open Music Observatory, the Slovak Comprehensive Music Database (https://hudobnadatabaza.sk/en/), and the Finno-Ugric Data Sharing Space (https://finnougric.net/), already demonstrate this approach in practice 1. The present feasibility study assesses the applicability of the Heritage Digital Twin Ontology (HDTO) draft on complex digital representations of both material and immaterial heritage. It also provides the conceptual and methodological background for three datasets: Livonian Music Heritage Digital Twins,Finno-Ugric Material Heritage Twins, and one cross-domain dataset extending the results of OpenMusE. As a synthesis of OpenMusE and a pilot for FinFAIR, this report provides: 1. A transparent account of how HDTO is operationalised in real-world data environments; and 2. A blueprint for transforming heterogeneous, multilingual heritage datasets into sustainable, semantically linked resources ready for integration into the European Collaborative Cloud for Cultural Heritage. The data curatorial and governance logic of our work, with some technical implementation aspects was presented on Wikidata Conf 20252. Impact Pathways Our planned FinFAIR implementation demonstrates how small and regionally rooted cultural institutions can contribute to Europe’s digital transformation on equal terms with large heritage organisations. It provides legal, organisational, semantic, and technical support for connecting the curatorial and collection-management work of national, regional, and village GLAM institutions, community archives, and private collectors. Through this collaboration, digital heritage is no longer confined to isolated catalogues but becomes part of an open, interoperable, and ethically governed ecosystem. The creation of Heritage Digital Twins allows cultural expressions to be represented in their full richness. Songs, stories, artefacts, and performances are linked into composite entities that combine sound, text, image, and context. This structure preserves curatorial depth 1Our definition here paraphrases (Curry 2020) and the Data Space for Cultural and Creative Industries report (EBU and Gaia-X 2022, p16): a “data [sharing] space” is an ecosystem for the exchange, processing, and provision of data between trusted partners. For the Open Music Observatory, see (Daniel Antal 2024). 2See references: (Commission et al. 2022; ECHOES Ontology Task Force 2025); conference presentation: (Daniel Antal 2025). 2 while making heritage data reusable across disciplines. It supports research in ethnomusicology, linguistics, and anthropology, while also giving communities direct access to resources that strengthen local identity and cultural continuity. Our methodological emphasis on rights-aware design ensures that every digital component can be reused safely and ethically. By applying the HDTO and ODRL frameworks, we aim to clarify authors’ rights, neighbouring rights, data privacy, and the CARE-based ethical norms surrounding archival recordings, photographs, and transcriptions—materials that have often remained inaccessible because of uncertain licensing. This approach supports a new generation of trustworthy heritage data, enabling both scientific inquiry and community creativity, including language re-learning and participatory cultural revival. For small ethnolinguistic communities such as the Livonians and Latgalians, the impact of the Finno-Ugric Data Sharing Space and the broader Open Music Observatory lies in visibility and participation. When traditional songs or artefacts become findable and audible online, they invite renewed engagement—singing, learning, and reinterpretation. Through collaboration with Wikimedia Eesti, these digital twins extend into the Wikimuseum environment, where regional museums and citizen curators can co-create multilingual exhibitions without moving physical objects. By combining open technologies with responsible governance, FinFAIR demonstrates how the principles of the European Collaborative Cloud for Cultural Heritage can be realised in practice. It links the technical innovation of the digital twin model with social and linguistic sustainability, showing how data infrastructures can support living heritage rather than merely document it. This feasibility study demonstrates how the Heritage Digital Twin Ontology (HDTO) can be practically applied to integrate tangible (recordings, notations, instruments) and intangible (songs, repertoires, performance traditions) aspects of cultural heritage. By linking these components into a single semantic structure, HDTO enables the reconstruction of cultural meaning that would otherwise remain fragmented across archives, institutions, and languages. The HDTO approach provides a valuable perspective on the challenges addressed by the Open Music Europe project, even though its primary focus is rights management rather than heritage management. HDTO distinguishes between the immaterial cultural object (a musical work) and its digital twin—a structured digital representation encompassing digitised scores, textual manifestations, and audio or video recordings. Creating such a composite object and linking it to the immaterial musical work parallels the workflow of rights management agencies that license recordings or scores of a given work. The key difference is that the digital twin requires more explicit and structured rights metadata than traditional heritage documentation. Combining heritage and rights-management data significantly increases the usability and impact of digital heritage: many sound and video recordings held by cultural institutions remain inaccessible precisely because their rights status is unclear or undocumented. 3 Demonstrator Datasets The demonstrator datasets were designed to test whether the principles of the data sharing space and the European Interoperability Framework can be applied to the early-stage Heritage Digital Twin Ontology (HDTO). Particular attention was given to the legal, organisational, and semantic challenges of constructing digital twins from materials originating in different institutions, collected through distinct curatorial pipelines, and governed by heterogeneous rights, duties, and permissions. These datasets act as proofs of concept showing how distributed heritage materials can be reconnected across linguistic, institutional, and legal boundaries while preserving community authority and ethical integrity. Livonian Music Heritage Digital Twin Dataset •Scope: Approximately 50 high-context records of Livonian music, with potential parallels from censored or restricted Hungarian recordings, curated according to the Heritage Digital Twin Ontology developed within ECHOES /ECCCH. •Focus: Demonstrates a community-centred, decolonising approach to metadata repair that provides ex post epistemic justice for displaced, marginalised, or silenced performers. It applies FAIR and CARE principles to ensure ethical and transparent rights management. •Function: Illustrates how HDTO restores semantic relationships among musical components—melody, lyrics, textual and notated manifestations, and recordings that carry distinct rights layers. It highlights recurring challenges of GDPR and CARE compliance in sound archives, especially where political transitions, institutional closures, or missing provenance data have disrupted legal continuity. •Data Sources: –Three collections of the Archives of Latvian Folklore, –the Hõimulõimed Finno-Ugric Song Dataset of 2 025 vocal recordings with lyrics in Livonian, Karelian, Veps, Sámi, Khanti-Mansi, Komi, Mari, Udmurt, Erzyan, Moksha, and Samoyedic languages 3, –open databases such as Discogs and MusicBrainz, and authority control via Wikidata,VIAF, and ISNI. Public and private sources are connected, including limited-access streaming platforms (e.g. Spotify), demonstrating how licensed and open materials can coexist within one rights-aware infrastructure. •Access: Via https:://finnugric.net or (Daniel Antal and Mester 2025) 3The Hõimulõimed Finno-Ugric Song Dataset was collected by Britt-Kathleen Mere, Bogáta Timár, Agatha Murro, and members of the Hõimulõimed cultural organisation between July 2023 and early 2025. The dataset was subsequently enriched and uploaded into the Finno-Ugric Data Sharing Space. 4 Finno-Ugric Material Heritage Digital Twins Dataset •Scope: Approximately 20 high-context records of Livonian vernacular buildings and associated artefacts curated according to the Heritage Digital Twin Ontology within ECHOES /ECCCH. •Focus: Demonstrates community-centred and decolonising metadata practices in documenting tangible heritage. Applies FAIR and CARE principles to balance open accessibility with local custodianship and cultural sensitivity. •Function: Shows how HDTO reconnects tangible components—a homestead, its constituent buildings, interior structures, and related ethnographic objects—with historical and contemporary photographs. It provides a comparative perspective on modelling tangible versus intangible aspects within a single digital-twin framework. •Data Sources: –The Ethnographic Open-Air Museum of Latvia, –the Ventspils Museum, and –original photographic documentation collected by the project team. •Access: Via https://finnougric.net/ or (Daniel Antal, Mester, and Pigozne 2025) Together, these demonstrators explore how diverse heritage materials—musical and architectural, intangible and tangible—can be curated through a shared ontology and governance model. They validate that Heritage Digital Twins can maintain semantic precision, multilingual accessibility, and long-term interoperability while respecting community ownership and ethical data sharing. Data Sources and Modelling Framework This section describes the data sources and ontological design principles applied within the Finno-Ugric Data Sharing Space (FUDSS), which forms part of the Open Music Observatory. The data come primarily from the Nordic and Baltic regions and from Finno-Ugric ethnolinguistic groups. Together, they serve as the empirical foundation for testing how tangible and intangible heritage can be represented as interoperable digital twins across community, institutional, and regional boundaries. Each dataset is semantically aligned to the Heritage Digital Twin Ontology (HDTO) classes (hdt:HC1–HC7) and properties (hdt:HP1–HP3,dcterms:license,prov:wasDerivedFrom), ensuring consistency with ECCCH ontological standards and long-term interoperability. Interoperability and Governance The project builds upon a decentralised data-sharing-space governance model developed through the OpenMusE Horizon Europe project and extended through collaborations with Wikimedia chapters. This prototype infrastructure already hosts several high-value 5 datasets used for methodological testing and feasibility analysis, establishing FUDSS as a functioning federated environment. Initial ingestion activities focused on methodological diversity, testing how a federated community infrastructure could interoperate with ECCCH,Europeana,OpenAIRE, and DARIAH 4. These early datasets remain open for reuse and further ingestion from Finnish, Estonian, and Hungarian national Finno-Ugric collections, as well as from NGOs representing ethnolinguistic communities. The technical feasibility phase concentrated on music as intangible heritage and dress history as tangible heritage. It produced a sample dataset and a methodological data paper based on early HDTO drafts. Consultation with domain experts and ontology developers broadened the scope to include built heritage, audiovisual materials, ethnographic artefacts, and still images (photographs, lithographs, prints). Although FinFAIR itself has limited resources for new curation, they are used strategically to produce an HDTO-validated dataset that provides a near-complete representation of Livonian heritage and a comparative dataset for the Latgalian community. Because the Livonian corpus is relatively small, the project aims for exceptionally high coverage—representing at least half of the known material and immaterial heritage of this group. This approach prioritises depth of modelling and rights clarity over raw data volume, ensuring that the demonstrators provide meaningful technical feedback to the ECCCH ontology developers. Open Modelling Questions and Possible HDTO Improvements Feasibility testing and dataset creation revealed a structural gap in HDTO v0.1 concerning the relationship between the conceptual digital twin (HC2_HeritageDigitalTwin) and its concrete digital components (HC5–HC8_DigitalRepresentation). Because HC2 is modelled as a subclass of E89_PropositionalObject and HC5–HC8 descend from crmdig:D1_DigitalObject (under E70_Thing), the two branches are formally disjoint in CIDOC-CRM. The published ontology therefore lacks a direct property connecting a digital twin to its digital surrogates, which can cause accidental mis-ranging when implementers reuse HP3_hasDigitalTwinComponent. To address this in practice, the FUDSS introduces a temporary bridging property: Proposed Curatorial Work Our feasibility study and the creation of a sample dataset demonstrated the advantages of the Heritage Digital Twin Ontology (HDTO) in settings with limited archival depth and fragmented documentation. The digital-twin approach supports not only metadata repair and enrichment but also the decolonisation of cultural records created under unequal historical conditions. For Finno-Ugric and Baltic minorities, documentation is often incomplete, dispersed, or filtered through linguistic and curatorial mediation. Creating digital twins in 4For interoperability references, see (Europeana 2017; Manghi et al. 2019; DARIAH-ERIC 2024). 6 these contexts means reassembling traces of a cultural entity across archives, media types, and languages. Even within a single institution, elements of a musical work—handwritten scores, printed editions, recordings—are frequently detached from one another. By linking these components into a composite digital-twin object, we follow the logic of the emerging Records in Contexts (RiC) model, which represents archival entities as dynamic networks rather than isolated entries. HDTO enables the construction of such semantic relationships, allowing us to represent a work within its historical and linguistic context. A single example illustrates both the practical and conceptual dimensions of this approach. While analysing a Latvian-language wedding song from Northern Courland, Šķiraties, rumaties, puķi vedu istabāi, we located a related variant with mismatched catalogue numbers. Further inspection revealed Es ar sav’ baleliņ, a phonetic transcription of the same melody recorded by ethnographer Emilis Melngailis in 1930 in Pitrags. As was common at the time, Melngailis transcribed oral materials by ear in Latvian orthography, inconsistently applying diacritics and masking key Livonian phonetic features. Although the original text cannot be fully reconstructed, fragments correspond closely to known Latvian lyrics of the same song. From this process, we created two digital twins: one for the Latvian-language variant (Šķiraties, rumaties, puķi vedu istabāi) and another for the Livonian version (Es ar sav’ baleliņ). Each carries its own linguistic and cultural context yet remains semantically linked to the other as a probable variant of the same musical work. This approach allows them to coexist as culturally distinct but interconnected heritage entities, avoiding ethnic essentialism by situating the song within shared Latvian, Livonian, and Finno-Ugric traditions. The example also reflects the layered colonial histories that shaped these data. Northern Courland passed through Russian, German, and Soviet rule, while the Livonians—a small Finno-Ugric minority—were largely assimilated by the early twentieth century. Their heritage was mostly documented by external ethnographers from Latvia, Finland, Estonia, and Hungary, whose transcriptions and classifications often obscured linguistic features and local meanings. To make such collections meaningful for contemporary Livonian speakers and researchers, it is necessary to reconstruct not only the metadata but also the epistemic conditions under which they were produced. The digital-twin approach enables this reconstruction by linking dispersed records across languages, archives, and political regimes into shared semantic structures. Within this framework, HDTO provides a means of representing heritage that acknowledges multilingualism, displacement, and asymmetrical authorship while restoring interpretative authority to the communities themselves. 7 Implementation Why Wikibase as an Architectural Choice Wikibase provides an effective bridge between curatorial workflows and semantic-web modelling. As an open, collaborative database, it allows curators and researchers to describe heritage materials in a human-readable form while keeping data machine-actionable through RDF export. This duality makes it ideal for projects balancing community participation, multilingual metadata, and interoperability with formal ontologies such as HDTO, CIDOCCRM, or schema.org. In contexts of limited archival depth and dispersed or diasporic heritage, strict ontology modelling can be too rigid for practical curation. Wikibase supports incremental enrichment: incomplete or conflicting data can coexist with qualified statements, references, and provenance notes, allowing curators to capture uncertainty without breaking semantic consistency. Its flexible data model permits the gradual introduction of domain ontologies—digital-twin classes, heritage-entity relations, and component hierarchies—without disrupting existing documentation. Architecturally, Wikibase serves as a bridging layer between descriptive archives and semantic infrastructures. It connects traditional cataloguing and fieldwork metadata with graph-based reasoning, while federated queries and reconciliation features enable linkage across institutions and languages. Within the Finno-Ugric Data Sharing Space, this approach allows complex, multilingual, and historically layered heritage to be represented transparently and sustainably: human curators maintain an accessible knowledge base, while automated exports produce HDTO-compliant datasets for long-term interoperability. Wikibase has proven effective for multilingual authority control (Bianchini, Bargioni, and Pellizzari di San Girolamo 2021; Fagerving 2023). The Finnish Wikibase pilot of the National Library of Finland demonstrated how it can connect museum, linguistic, and archival records while supporting community participation (National Library of Finland 2021). Similar experiences in minority-language contexts—such as the Võro User Group (Wikimedians of Võro language User Group 2025) and Wikimedia Norge’s Sámi knowledge project (Wikimedia Norge 2020)—show that Wikibase can combine semantic interoperability with respect for knowledge sovereignty and multilingual representation. These precedents confirm that Wikibase can act as a shared curatorial platform for minority and diasporic heritage without imposing a single institutional metadata regime. By aligning the Wikibase data model with the Heritage Digital Twin Ontology (HDTO), curated materials in our knowledge graph can be exported directly into the semantic infrastructure of the European Collaborative Cloud for Cultural Heritage (ECCCH). This export layer transforms each composite Wikibase record into a structured HDTO dataset, linking it to wider European data spaces through persistent identifiers, multilingual vocabularies, and standardised provenance. In this way, Wikibase functions not only as a curatorial interface but also as an enabling architecture for open, interoperable cultural-heritage data within the evolving European research ecosystem. 8 Data Model Implementation Our knowledge graph represents cultural-heritage entities and their digital counterparts according to the Heritage Digital Twin Ontology (HDTO) used by the European Collaborative Cloud for Cultural Heritage (ECCCH). In HDTO v0.1, the class HC1_HeritageEntity denotes any real-world object—tangible or intangible—valued for its cultural, social, or scientific significance. It is a subclass of crm:E70_Thing and covers physical, immaterial, or born-digital heritage recognised as having enduring importance. Individual events are not HC1 instances but may be associated through event types or patterns. Each real-world heritage entity—for instance, a song, manuscript, or artefact—is represented as an instance of hdt:HC1_HeritageEntity. These entities are not stored directly in Wikibase but generated dynamically during export to an HDTO-compliant dataset. Digital twins and their heritage entities are linked through hdt:HP1_hasDigitalTwin (inverse hdt:HP1i_isDigitalTwinOf), which asserts that a digital object documents a realworld entity. Within each composite twin we distinguish between – the main digital twin (the Wikibase item itself), and – its digital components, corresponding to individual files or modalities such as lyrics, audio, images, or scores. Material Assets The Dēliņi Homestead, constructed between 1819 and the early twentieth century, forms a distinctive three-courtyard layout (fudss:Q5582). Originally located in Lūžņa on the Livonian coast, it was relocated to the Latvian Ethnographic Open-Air Museum in 1971. :Delini_Homestead a `hdt:HC1_HeritageEntity` ; rdfs:label "Dēliņi Homestead"@en ; `hdt:HP15_isHeritageOf` :Livonian_Community . A combined barn-shed built around 1870 is modelled as a hdt:HC3_TangibleAspect of the homestead. :Delini_BarnShed a `hdt:HC3_TangibleAspect` ; `crm:P46_is_composed_of` :Delini_Ramp ; `hdt:HP1_hasDigitalTwin` :Delini_BarnShed_Twin . Three photographs document the barn-shed; each image is a separate creative work with its own licence. 9 an Estonian-only thesaurus. Building full mappings between AAT and muis.ee would itself constitute a large, independent project. Our approach combines Wikidata’s pragmatic “good-enough” mapping with the eXtreme Design methodology used in the Polifonia ontology family (Blomqvist, Hammar, and Presutti 2016; Berardinis et al. 2023). This method reuses semantic fragments to address explicit competency questions—for example, how to represent garment types or musical forms—without needing to localise entire thesauri. We follow AAT’s localisation guidelines only in narrow areas where we have sufficient searchable objects and digital twins. Most Latgalian and Livonian garment collections, for instance, date from the late nineteenth century. We therefore harmonised only the vocabulary for that period’s peasant dress across Estonian, Finnish, Latvian, and AAT schemas. AAT’s upper-level structure (“shirt”, “upper-body garment”) serves as a scaffold for culturally specific subclasses such as kitasnik (dress) and kuhiksõlg (Seto conical brooch). Although AAT does not yet cover all FinnoUgric material culture, it provides a workable structure for cross-institutional alignment and diasporic item discovery. Where AAT coverage is insufficient—especially for intangible heritage like music genres or ritual practices—we extend alignment through Wikidata property sets and local SKOS concept schemes within the Finno-Ugric Data Sharing Space (FUDSS). For linguistic and cultural concepts unique to small Finno-Ugric communities, we apply OntoLex-Lemon to build lexeme-based thesauri capturing variant spellings, etymologies, and semantic overlaps among related languages (Livonian, Võro, Seto, Mari). Each lexical entry is mapped to AAT or Wikidata equivalents (skos:exactMatch,skos:closeMatch, skos:broadMatch) to enable multilingual query expansion and culturally specific representation. Spatial data pose similar challenges, as historical sources often use multilingual or obsolete place names. Geographical Namespaces and Spatial Authorities Place names are a major source of inconsistency across archival sources. Historical materials often employ outdated or colonial-era terminology, while modern boundaries differ substantially. FinFAIR addresses this through a dual-namespace model: •Contemporary entities: normalised via GeoNames and Wikidata geospatial ontologies. •Historical entities: preserved as custom SKOS concepts or FUDSS-local authorities derived from archival inventories. 16 We built a multilingual reconciliation layer—a gazetteer—linking historical and modern names (e.g. Petseri, Pechory, Petschur) under canonical identifiers. Beyond standard geocoding, our approach integrates multi-script, multi-ethnic, and historical variants verified by Finno-Ugric scholars. The Livonian coast serves as our prototype. Using the Livonian Place Name Catalogue (Ernštreits 2020; Ernštreits, Šuvcāne, and Damberg 2024), we compiled names across Livonian, Latvian, Estonian, Finnish, Hungarian, Russian, and Baltic German. This multilingual layering improved recall of heritage materials in foreign or colonial catalogues. Even within a dozen villages, orthographic variation is vast—e.g. Irē (Livonian) vs. Irben or Gross-Irben (German). Soviet-era catalogues standardised Latvian forms; imperial sources retained German exonyms. These linguistic and political layers illustrate why rigid topdown normalisation is impractical. We therefore published a prototype gazetteer linking (Ernštreits 2020; Ernštreits, Šuvcāne, and Damberg 2024; Daniel Antal, Pigozne, and Mester 2025) to multilingual identifiers. It connects surface forms of Livonian, Estonian, Russian, and German with standardised Livonian and Latvian names. This enabled the discovery of Livonian garments in Finnish and Estonian collections previously unknown to Latvian researchers. Each location node stores both current and historical identifiers, allowing queries by modern coordinates or original ethnographic appellations. For example, “Kolka” (modern Latvian) and “Kūolka” (Livonian) are distinct labels of the same entity—preserving linguistic distinction while maintaining referential unity. Authority Files for Persons, Organisations, and Works Authority control ensures continuity between FinFAIR datasets and the wider European infrastructure. It links people, institutions, and works across repositories, enabling both automated interoperability and human validation. •Persons: aligned through VIAF,ISNI, and Wikidata; when absent, internal IDs with open crosswalks. •Institutions: identified via ROR and Wikidata, ensuring compatibility with OpenAIRE and EOSC. •Works: linked to MusicBrainz,Wikidata,GLAM Wiki Works, or national library records. •Sound recordings: identified by ISRC and, where accessible, platform identifiers (Spotify, Deezer, YouTube), ensuring discoverability and lawful reuse. 17 Ontological Alignment HDTO functions as the central semantic layer, extended and cross-referenced with established ontologies: Domain Ontology / Schema Purpose Tangible heritage CIDOC-CRM 7.3 Provenance, physical description Archival context Records in Context (RiC-O) Archival hierarchy and ISAD(G) compatibility Bibliographic data DCTERMS, BIBFRAME Publication and citation Music and sound DDEX, Polifonia Ontologies Rights, instrumentation, performer roles Language and lexicon OntoLex-Lemon Linguistic data and lexical thesauri Rights management ODRL Structured rights policies for digital components Mappings between these ontologies are defined through owl:equivalentClass and owl:equivalentProperty within the FUDSS namespace. This ensures that exported datasets merge seamlessly with ECCCH and Europeana ingestion pipelines. Because of our architecture, HDTO operates both within the Wikibase Data Model and as a formal ontology. As HDTO derives from CIDOC-CRM, already mapped to Wikibase (Wikimedia Foundation 2025; Pfeiffer and Gayo 2021; Kesäniemi, Koho, and Hyvönen 2022), we successfully aligned version 0.1 to produce a demonstration dataset. Following the pragmatic logic applied to vocabularies, we do not aim to map DCTERMS, RiC-O, or CIDOC-CRM in their entirety. Instead, using eXtreme Design, we formulate competency questions tailored to our institutional data sources—those in FinFAIR and the Finnish, Estonian, and Hungarian repositories—to ensure interoperability across bibliographic, archival, and museum records. Our alignment extends to DDEX, the music-industry metadata standard. Aware of metadata bias and “ontological hijacking,” our mapping remains minimal and principled: ensuring coherence between title and name headings (for works and recordings) and between time and place appellations used in CIDOC-derived systems and music data sources. The feasibility dataset confirms this high-level interoperability in practice. Multilingual Data Representation All human-readable labels (rdfs:label,skos:prefLabel) are provided in at least two languages: English and the original language of the source (Latvian, Estonian, Finnish, Hungarian, or minority languages such as Livonian or Võro). Transliterations follow ISO 9:1995 and established national orthographic standards for non-Latin scripts. This ensures 18 that ECCCH harvesters can retrieve English-language metadata for interoperability, while community users can query and display data in their native scripts. Multilingual representation is not only a usability feature but a core part of FinFAIR’s ethical and epistemic framework. It recognises linguistic diversity as a form of intangible heritage and ensures that endangered or minority languages remain visible in digital infrastructures. Every dataset therefore carries both internationally legible and locally meaningful identifiers, balancing machine readability with cultural specificity. At a later stage, we plan to introduce explicit lexicographical modelling through the Lexeme extension of Wikidata and Wikibase, which implements the OntoLex-Lemon ontology. This will allow the project to represent linguistic variation, etymology, and dialectal features more precisely and to interlink cultural-heritage data with emerging European Language Data Space resources. Rights Management Rights management within the FinFAIR framework is designed to be transparent, interoperable, and machine-readable. It combines the Heritage Digital Twin Ontology (HDTO) with practical extensions in Wikibase that allow curators to express not only licences but also contextual usage rules derived from the Open Digital Rights Language (ODRL). While large digital-heritage platforms such as Europeana or DARIAH rely mainly on singleURI licence models (edm:rights,dcterms:license), FinFAIR adds an optional layer of structured use policies to handle complex, multimodal works such as music or film, where overlapping rights coexist. Personal Data and GDPR Every hdt:HC1_HeritageEntity or hdt:HC2_HeritageDigitalTwin may link to living agents—performers, collectors, researchers, donors, or other data subjects—through dcterms:creator,prov:wasAttributedTo, or crm:P14_carried_out_by. Under the General Data Protection Regulation (GDPR), these names constitute personal data. Before publication or linkage, consent and attribution are evaluated: •Attribution — if the person consents to credit, we use cc:attributionName. •Pseudonymisation — if consent is uncertain, we use neutral labels, e.g. “Field recording performer, 1962”. •Anonymity — if the person requests privacy, all identifying information is withheld. • The selected attribution mode determines the applicable licence. For example, CC-BY assumes credit, while pseudonymised or anonymous materials default to CC0,BY-NC, or a corresponding rightsstatements.org URI. 19 For older archival materials, FinFAIR follows Europeana’s approach: if the data subjects are certainly deceased (typically pre-1925), personal-data restrictions no longer apply. Simple Case: One Licence per File For non-composite objects—photographs, scans, or plain-text documents—the Europeana Data Model (EDM) edm:rights field provides sufficient legal clarity. FinFAIR implements this directly via the Wikibase property use rights, which corresponds to dcterms:license and edm:rights. Even these “simple” cases are dynamic. A copyright protection term may expire, making a less accessible item fully reusable under the public domain. Rightsholders may change their minds, and give a more permissive license, for example, change from CC BY-NC-ND to CC-BY-SA, allowing commercial use and derivate works. Type of object HDTO class Typical examples Digital text hdt:HC6_DigitalDocumentmetadata exports, lyrics, PDFs Digital image hdt:HC7_DigitalVisualObjectphotographs, scans, blueprints Simple audio file hdt:HC5_DigitalRepresentationsingle, unambiguous recordings Each digital component carries: - use rights → a URI pointing to a standard licence (e.g. CC-BY, CC0); - optional provenance (dcterms:rightsHolder,dcterms:accessRights,dcterms:provenance). 20 The composite digital twin (HC2) aggregates multiple components but does not override their licences. When reused, the most restrictive licence among components applies. Wikibase Rights Extensions To align HDTO, EDM, and Wikibase, FinFAIR defines two complementary properties and one supporting class. Property: use rights (wd:P486) Field Description Label use rights Description Links a digital object to a URI identifying the applicable licence or rights statement (e.g. rightsstatements.org, Creative Commons). Aliases licence; rights; edm:rights; dcmi:license; datacite:rights Datatype External identifier (URI) Equivalent properties dcterms:license,edm:rights,datacite:rights Full definition https://reprexbase.eu/fu/Special: EntityData/P486.ttl Example: •Dēliņi homestead detail: combined barn-shed (detail photo) use rights:https://creativecommons.org/licenses/by-sa/4.0/deed.en •The Livonian Community House in Mazirbe (photograph) use rights:https://creativecommons.org/licenses/by/4.0/deed.en This property provides one-to-one interoperability with both Europeana and research repositories using DataCite Property: use policy (wd:P487) Field Description Label use policy Description Links a digital object to a local rights or usage policy item within the same Wikibase instance (e.g. “Listen on Spotify”). Used when the applicable terms are described internally rather than via an external URI. 21 Field Description Aliases has policy; user rights policy Datatype Wikibase item (Near) Equivalent class odrl:Policy Full definition https://reprexbase.eu/fu/Special: EntityData/P487.ttl Class: use policy (wd:Q5606) Field Description Label use policy Description A simplified representation of odrl:Policy within Wikibase. Describes how a digital object may be accessed, used, or shared under specific conditions. Instances act as human-readable policy nodes. Full definition https://reprexbase.eu/fu/Special: EntityData/Q5606.ttl Example instance of this class: Q5608 – Listen on Spotify You may listen to this recording via the Spotify web or mobile player, either for free or with a subscription, under Spotify’s End-User Licence. This simplified approach to ODRL was first introduced in our WikidataCon 2025 presentation, where we demonstrated how Wikibase can represent policy semantics (“play”, “reproduce”, “not download”) without external RDF frameworks. Complex Case: Composite Rights Multimodal cultural objects, such as sound recordings, often involve overlapping rights: composer, lyricist, performer, producer, and distributor. In such cases, each digital representation (HC5) carries its own use rights and use policy. 22 Example: Tšitšōrlinki, tšitšōrlinki •Archival recording — accessible under CC-BY-NC, for non-commercial research. use policy: “Non-commercial analysis and reproduction permitted.” •Spotify release — available for streaming only under Spotify’s end-user licence. use policy: “Listen on Spotify.” These are linked within a single digital twin (HC2), which aggregates but does not merge or override rights. 23 Aggregators such as Europeana can display the most restrictive licence, while research or educational users can act according to the specified use policy. Summary 1. GDPR compliance — screen linked agents, pseudonymise or anonymise where required. 2. Simple case — one use rights URI per file, equivalent to edm:rights. 3. Composite case — multiple components, each with its own use rights and use policy. 4. Interoperability — full equivalence with EDM, DCTERMS, and ODRL. 5. Scalability — lightweight for images and texts, yet expressive for music and audiovisual works. This model provides a sustainable bridge between cultural heritage and research data ecosystems, ensuring that digital twins can circulate legally and ethically within the European Collaborative Cloud for Cultural Heritage. References Antal, Daniel. 2024. “Open Music Observatory.” Digital Music Observatory. https://doi. org/10.5281/zenodo.16539570. ———. 2025. “Wikibase as a Data Sharing Space: Connecting Rights, Communities, and GLAM Through Federated Infrastructures. Presentation on Wikidata Conf 2025.” Open Music Observatory. https://doi.org/10.5281/zenodo.17496740. 24 Antal, Daniel, and Anna Márta Mester. 2025. “Finno-Ugric Material Heritage Digital Twins Dataset.” Open Music Observatory. https://doi.org/10.5281/zenodo.17492052. Antal, Daniel, Anna Márta Mester, and Ieva Pigozne. 2025. “Finno-Ugric Material Heritage Digital Twins Dataset.” Finno-Ugric Data Sharing Space. https://doi.org/10.5281/ zenodo.17492092. Antal, Daniel, Ieva Pigozne, and Anna Márta Mester. 2025. “Remapping the Livonian Coast: A Multilingual Gazetteer of the Settlements of Northern Kurzeme.” Finno-Ugric Data Sharing Space. https://doi.org/10.5281/zenodo.15668712. Antal, Dániel, Ieva Pigozne, Kata Gábor, and Bogáta Tímár. 2025. “Federating Open Knowledge Graphs: Finno-Ugric Cultural Heritage and Semantic Infrastructure Facilitated Through Wikibase [Preprint].” In Proceedings of the Digital Humanities in the Nordic and Baltic Countries (DHNB) Conference. Vol. 7. CEUR-WS. Archives Expert Group on Archival Description, International Council on. 2023. “Records in Contexts–Conceptual Model. Version 1.0.” International Council on Archives. https: //www.ica.org/app/uploads/2023/12/RiC-CM-1.0.pdf. Bekiari, Chryssoula, George Bruseke, Erin Canning, Martin Doerr, Philippe Michon, Christian-Emil Ore, Stephen Stead, and Velios Athanasios, eds. 2024. “Definition of the CIDOC Conceptual Reference Model.” CIDOC CRM Special Interest Group. https://www.cidoc-crm.org/sites/default/files/cidoc_crm_version_7.2.4.pdf. Berardinis, Jacopo de, Valentina Anita Carriero, Nitisha Jain, Nicolas Lazzari, Albert Meroño-Peñuela, Andrea Poltronieri, and Valentina Presutti. 2023. “The Polifonia Ontology Network: Building a Semantic Backbone for Musical Heritage.” In The Semantic Web – ISWC 2023, edited by Terry R. Payne, Valentina Presutti, Guilin Qi, María Poveda-Villalón, Giorgos Stoilos, Laura Hollink, Zoi Kaoudi, Gong Cheng, and Juanzi Li, 302–22. Lecture Notes in Computer Science. Cham: Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-47243-5_17. Bianchini, Carlo, Stefano Bargioni, and Camillo Carlo Pellizzari di San Girolamo. 2021. “Beyond VIAF Wikidata as a Complementary Tool for Authority Control in Libraries.” Information Technology and Libraries 40 (2). https://doi.org/10.6017/ital.v40i2.12959. Blomqvist, Eva, Karl Hammar, and Valentina Presutti. 2016. “Engineering Ontologies with Patterns – the eXtreme Design Methodology.” In Ontology Engineering with Ontology Design Patterns, 23–50. IOS Press. https://doi.org/10.3233/978-1-61499-676-7-23. Commission, European, Directorate-General for Research, Innovation, P. Brunet, L. De Luca, E. Hyvönen, A. Joffres, et al. 2022. Report on a European Collaborative Cloud for Cultural Heritage – Ex – Ante Impact Assessment. Publications Office of the European Union. https://doi.org/doi/10.2777/64014. Core, Dublin. 2020. “DCMI Metadata Terms.” http://dublincore.org/specifications/dublincore/dcmi-terms/2020-01-20/. Curry, Edward. 2020. “Dataspaces: Fundamentals, Principles, and Techniques.” In RealTime Linked Dataspaces: Enabling Data Ecosystems for Intelligent Systems, 45–62. Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030-296650_3. DARIAH-ERIC. 2024. “Digital Research Infrastructure for the Arts and Humanities (DARIAH) – Infrastructure Overview.” https://www.dariah.eu/.https: //www.dariah.eu/. EBU, and Gaia-X. 2022. “Dataspace for Cultural and Creative Industries. Position Paper. 25