ENEOLI Wikibase: A collaborative working platform for the European Network on Lexical Innovation

ENEOLI Wikibase: A collaborative working platform for the European Network on Lexical Innovation David Lindemann1, Ana Salgado2 1EHU University of the Basque Country, Unibertsitatearen ibilbidea 5, 01006 Vitoria-Gasteiz (Spain) 2University of Porto, Centre for Linguistics of the University of Porto, Porto (Portugal) E-mail: [email protected], [email protected] Abstract This paper presents the conceptual basis and practical implementation of a Wikibase instance as framework for collaborative workflows, developed as part of the COST action European Network on Lexical Innovation (CA22126). The platform supports the joint creation and publication of three interlinked resources, all integrated into a single knowledge graph: (1) NeoCorpus, a curated repository of bibliographic records and full text collection of the field of Neology; (2) NeoVoc, a multilingual metaterminological vocabulary designed to model, structure, and harmonise the conceptual and terminological frameworks used in research on neology and lexical innovation; and (3) a collection of neologisms, which documents and describes emergent lexical items across various European languages. Wikibase, a free and open-source software platform, provides a generic and flexible environment for collaboratively editing and publishing multilingual, cross-domain descriptions of concepts and lexical units as FAIR Linked Data. The paper explains how resources developed by different ENEOLI Working Groups are made interoperable and linked to each other on the Wikibase instance, describing workflows and preliminary results. We believe that our choice for Wikibase as infrastructure solution for a collaborative Knowledge Graph, here for the domain of Lexical Innovation, may serve as reference for similar endeavours. Keywords: Lexical Innovation, Neology, Neologisms, Knowledge Graph, Wikibase 1. Introduction Lexical innovation1— understood both as a field of study and as a linguistic process — refers to the creation of new lexical units used to designate emerging concepts, entities, and socio-discursive realities. As a subfield of linguistics, lexical innovation investigates the mechanisms, contexts, and motivations that underlie the creation, diffusion, and stabilisation of new lexical items within a language. It manifests in different ways, encompassing a range of processes and phenomena through which new lexical material enters the linguistic system. Among these, neology2— the formation and integration of neologisms — represents one of the most productive and visible expressions of lexical innovation. The latter may thus be regarded as a broader domain encompassing interrelated areas such as neology, morphology, lexicography, and others. Within this framework, neology focuses more specifically on the identification, description, and analysis of new words and meanings in contemporary use. Whether through coinage, borrowing, derivation, compounding, or semantic shift, the continuous renewal of a language’s lexicon reflects the speakers’ need to designate, structure, and communicate novel experiences and ideas. This process is therefore deeply intertwined with cultural, scientific, and technological developments. One of the most pressing issues in this area of research is the lack of a shared, structured metaterminology3for describing lexical innovation (Tallarico et al., 2024). Across different linguistic traditions, terms such as blending,loanword,splinter or mot-valise are defined in divergent — and at times contradictory — ways. A pertinent example is the term compocation, proposed by Cusin-Berche (2003) within a Francophone theoretical 1See https://eneoli.wikibase.cloud/entity/L2263. 2See https://eneoli.wikibase.cloud/entity/L1057. 3Metaterminology refers to the set of terms and concepts that form a specialised metalanguage used to describe and categorise terminological objects, ensuring their acceptability, coherence, and functionality within a given conceptual system (cf., Salgado et al., 2025). framework. The term was coined to describe a specific type of lexical blending in which there is no segmental overlap — in contrast to prototypical mots-valises, which are defined by the phonological fusion of shared elements. Compocation can only be considered an autonomous concept if one regards the mot-valise with homophonous overlap as the prototype of the category. In other languages, however, such a distinction is generally not deemed necessary. Instead, the semantic scope of established terms — such as blending in English — is simply extended to cover a wider range of formations. These divergences hinder comparative research, both within and across language communities, and undermine the cumulative value of lexicographic and terminological efforts. The need for a harmonised, multilingual, and conceptually coherent vocabulary has therefore become increasingly evident in recent years. On the other hand, the study of lexical innovation — and, more specifically, of neologisms — presents significant methodological challenges. Neologisms are often context-dependent, short-lived, and variably recognised across communities and languages. A neologism may emerge suddenly and vanish without trace, or it may gradually gain traction through sustained usage in specialised or popular domains. Tracking such dynamics demands flexible models that accommodate variation in salience, frequency, and status over time. Advances in corpus linguistics, digital lexicography, and knowledge engineering have provided powerful tools for documenting and modelling lexical change. Projects rooted in the digital humanities increasingly rely on Semantic Web technologies — Linked Open Data (LOD) — to structure, interconnect, and enrich lexical and terminological data across domains and languages. These developments facilitate the collaborative production of FAIR (Findable, Accessible, Interoperable, Reusable) resources and enable sophisticated forms of linguistic analysis, including the tracing of conceptual evolution, cross-linguistic comparison, and discourse-sensitive annotation. In this context, the European Network on Lexical Innovation (ENEOLI, CA22126) has created the ENEOLI Wikibase platform,4a collaboratively editable infrastructure deployed to support and interlink three central resources: 1. NeoCorpus: a curated collection of bibliographic records and full texts relevant to the study of lexical innovation; 2. NeoVoc: a multilingual metaterminological vocabulary designed to model and harmonise the conceptual tools of lexical innovation research; 3. Neologisms Collection: an evolving structured database documenting new lexical items across a range of languages. Wikibase,5the open-source software that also underlies Wikidata,6a very large open Knowledge Graph, as infrastructure solution offers a flexible and transparent environment for collaboratively modelling and publishing structured data describing lexical and conceptual phenomena. The integration of the aforementioned three resources into a knowledge graph enables researchers to curate Linked Data across different levels of analysis — 4For more information about the platform, user documentation, and access to content visualizations through SPARQL database queries, refer to the main page of ENEOLI Wikibase at https://eneoli. wikibase.cloud. 5See https://wikiba.se. 6See https://www.wikidata.org. from bibliographic metadata to terminological concepts and lexical entries — and also to query the so created dataset, and to obtain data-driven evidence related to their research questions. 2. NeoCorpus NeoCorpus is a collection of scientific articles dedicated to lexical innovation. It brings together publication metadata and full texts of both historical and contemporary works — from foundational theoretical texts to recent case studies — covering multiple languages, traditions, and methodological approaches. The collection and organisation of the bibliographical records and the corresponding full text documents are managed collaboratively using Zotero,7a widely used cloud-based reference management system. A Zotero group library8enables ENEOLI members to collect, curate, and categorise bibliographic records and store full texts relevant to the field. Publication metadata are publicly available from the platform in numerous formats for citation purposes, while the full text documents attached to the bibliographic records stay only available to ENEOLI members. Each record undergoes a validation process to ensure consistency, accuracy, and relevance. This includes verifying metadata such as authorship and other publication metadata relevant for identification and citation. Only validated bibliographical records are selected for integration into the ENEOLI Wikibase. The migration from Zotero to Wikibase9involves a process of semantic enrichment: literal strings (e.g., author names, journal titles, or publication languages) are reconciled against ontological entities. For example, authors are associated with persistent identifiers (Wikidata, ORCID, VIAF), which enables an enrichment of Wikibase items describing authors with additional statements stemming from these sources,10 and which avoids duplicate entries. Such author disambiguation, a unique identifier for every authoring person, is also essential for database queries involving authors and metadata of their publications, namely, as we will see, the neological terms they use in them. In a Wikibase, that identifier will always be a URI, consisting of its base URL, and a numeral identifier. Content of some other Zotero fields — such as publication year, document type, language, and permanent identifiers (DOI, ISSN, ISBN, handle) — are also represented in Wikibase as URI, enabling more precise and flexible queries. A key innovation of NeoCorpus lies in its integration with other ENEOLI resources. As explained in section 3.1, bibliographic entries are enriched with information about NeoVoc terms (and, possibly, items in the ENEOLI Wikibase Neologisms Collection) occurring in the text, together with their absolute frequency in that text. This interconnected structure transforms NeoCorpus from a static list of references into a dynamic, queryable environment that supports conceptual mapping, bibliometric exploration, and discourse-based analysis of lexical innovation. 7See https://zotero.org. 8The ENEOLI group library is accessible at https://www.zotero.org/groups/5449136/neocorpus. 9For the technical setup, see https://marketplace.sshopencloud.eu/workflow/P0siWJ. 10 The enrichment can be achieved by copying data from a source like Wikidata, or by specifying federated database queries, i.e. combine content from more than one database in a query result. In summary, NeoCorpus illustrates how the collaborative curation of a Linked Data collection can produce a robust foundation for interdisciplinary research. It also highlights the potential of connecting free software solutions to each other, namely Zotero with Wikibase, to produce FAIR research data. 3. NeoVoc This component of ENEOLI Wikibase aims at providing a multilingual terminological and conceptual framework for the study of lexical innovation. That is, instead of deciding for a lemma-based specialised dictionary of the domain or for a concept-based terminological resource, the proposed graph structure allows to have both: conceptual entries, where information pertaining to the concept is attached, and lexical entries, where the term is described. NeoVoc addresses a critical gap in lexical innovation research: the absence of a coherent and interoperable metaterminology capable of supporting comparative, multilingual, and interdisciplinary analysis. NeoVoc brings together diverse terminological traditions by combining two complementary approaches to vocabulary construction: •A ‘bottom-up’ text-based terminological approach—where concept entries are created from terms extracted from the French section of NeoCorpus. These include terms commonly used in academic and applied discussions of lexical innovation, particularly in the Francophone tradition, which has a long history of theorising lexical innovation. This tradition is shaped by the work of authors such as Jean-François Sablayrolles (Sablayrolles (2019)) and Bernard Quemada (Quemada (1967); Quemada (1971)), and is well represented in specialised publication collections like Neologica and Terminologies nouvelles. •A ‘top-down’ theory-driven approach, grounded in the conceptual matrix proposed by Sablayrolles (2019). This framework maps the various processes involved in lexical innovation — such as composition,siglaison (acronym formation), and troncation (truncation) — and is now systematically complemented with the corresponding result types (neologism types), such as compound,acronym, and clipping, as well as others like mot-valise or blend. As explained, NeoVoc describes conceptual entries, on the one hand, and lexical entries, on the other. Each lexical sense is linked to its corresponding conceptual entry. This basic modeling follows recent approaches in Terminology as Linked Data (Bosque-Gil et al., 2015) involving the Ontolex-Lemon vocabulary (McCrae et al., 2017), which is pre-set for the description of lexical entries in a Wikibase (for details, see also Lindemann, 2025). For NeoVoc conceptual entries, members of the network collaboratively provide: •multilingual designations (terms denoting the concept) in their respective working languages; •glosses (Salgado et al., 2025), i.e., concise explanations of the essential distinguishing features of a concept to support disambiguation, in their working languages; •conceptual relations (e.g., subclass of ) that allow for the creation of a structured conceptual map. In order to speed up the editing process, NeoVoc concepts have been mapped, where possible, to Wikidata items, so that the following information, if present in Wikidata, could be imported as draft data to be validated: •Multilingual Wikidata “preferred labels” for the concept, to draft designations; •Multilingual Wikidata “descriptions” for the concept, to draft glosses; •Weblinks to multilingual Wikipedia articles about the concept. In regular iterations, for each equivalent designation (understood as a term candidate, for which attestations will be searched for in the NeoCorpus texts of that language) marked by a collaborator as validated,11 a script creates a lexical entry, with the validated designation as lemma. A sense object is attached to that entry, and linked to the denoted concept. In case a lexical entry for the same language with the same lemma and the same lexical category is already described, an additional sense object is created for that entry. While the inner modeling of a lexical entry in any Wikibase follows the core of the Ontolex-Lemon model (Lindemann, 2025), also the model of the ontology-lexicon interface follows Ontolex recommendations (Bosque-Gil et al., 2015). After creation, members of the network provide for lexical entries: •Applicable linguistic descriptions such as the grammatical gender of the term; •Links to external resources where the term is described; •If applicable, links to an etymon (in another language), or a derivation source (in the same language) and to objects representing lexical innovation processes involved in the creation of the term, as well as to objects describing neologism types (e.g., anglicism). This dual model of conceptual and lexical resources in NeoVoc therefore does not merely constitute a flat list of terminological variants across languages. Instead, it establishes conceptual equivalence between terms (understood as lexeme senses) by shared ontological references, ensuring that each term maps onto a concept. This allows for a more precise comparison of how different linguistic communities describe and analyse lexical innovation, and helps to identify cases of conceptual divergence or partial overlap. From a technical perspective, the NeoVoc editing workflow benefits from relevant features of the Wikibase software. Each concept or lexical entry is stored as a distinct entity with a unique permanent identifier. Edits performed by the network members and by automated scripts are stored together with the entity data in the edit history, which remains prepared for being reviewed, and, if regarded necessary, reverted. In sum, NeoVoc plays a foundational role in the ENEOLI ecosystem. It offers not only a shared vocabulary for researchers working on lexical innovation, but also a structured and extensible conceptual model that underpins the semantic integration of bibliographic and terminological data, and collections of neologisms. As such, it contributes both to the internal coherence of the ENEOLI platform and, in general, to the development of robust linguistic datasets, reusable in further research. 11 The explicit annotation of the author of the validation makes the process transparent, since it enables querying for validation statistics. The interconnection between NeoCorpus and NeoVoc is central to the platform’s design. The following section outlines how these two components are semantically integrated to support concept-based corpus querying and multilingual term alignment. 3.1 Integration of NeoVoc and NeoCorpus The semantic integration of both resources begins with the annotation of bibliographic records in NeoCorpus using lexical entries from NeoVoc. A workflow, adapted from previous research (Lindemann, 2021), identifies occurrences of NeoVoc terms in the lemmatised full text of articles, and, thanks to the existing alignment of NeoVoc terms in different languages with NeoVoc concept entries, enables to extend a term use analysis to a crosslingual analysis of concepts in use. This process enables the enrichment of bibliographic metadata with terminological and conceptual information, allowing users to query the corpus according to: •terms denoting specific neological concepts; •terms preferred by authors or clusters of authors; •languages or publication periods in which certain terms are attested; •co-occurrence patterns of terms and concepts within or across texts. While NeoVoc enables the structured description of metaterminological concepts, a further layer of modelling is required to capture the actual neologisms emerging in diverse linguistic contexts. The next section introduces the Neologisms Collection, which complements NeoCorpus and NeoVoc by focusing on newly coined lexical items. 4. Neologisms Collection Several task groups within the ENEOLI network are dedicated to the description of neologisms, that is, newly emerging lexical items—across a wide range of European languages. The ENEOLI Wikibase provides a collaborative platform for this work, supported by the conceptual framework established by NeoVoc, which offers a controlled set of annotation values. The data model and editorial workflow used for the Neologisms Collection are closely aligned with those applied for the construction of the NeoVoc metaterminology. Initially designed and tested in the context of NeoVoc, this approach has since been adopted by task groups focusing specifically on neologism documentation. Drawing on the insights gained from these pilot studies, we are working towards a fully documented, generic workflow for the description of neologisms within the ENEOLI Wikibase — one that can be reused and adapted by other groups in the network. At present, the Neologisms Collection comprises two main subsets: 1) Neology and Gender Equality (task 3.6); 2) 1st ENEOLI Training School – Lisbon 2025. The description of each neologism includes the following data categories, to be attached as Wikibase statements: •neologism type (e.g., anglicism) •innovation process involved in its formation (e.g., blending,derivation,borrowing); •lexical category and grammatical gender; •usage attestations and their sources; •links to external resources describing the item. As seen for NeoVoc lexical entries, a neologism entry word sense object can be linked to a conceptual entry as ontological reference. Such conceptual entries might be defined, following Ontolex-Lemon, as Lexical Concept, i.e., similar to Wordnet Synsets, which would group synonym senses with the same part of speech. In the case of nominal concepts, that referenced item, in turn, can be linked to or enriched with information present in Wikidata, where concepts associated to parts of speech other than nouns are not represented (Nielsen, 2020). A distinctive feature of the ENEOLI approach is a discourse-based and context-sensitive definition of neology. Rather than relying solely on chronological criteria, ENEOLI considers a lexical item to be a result of a lexical innovation process (i.e. a neologism), if it continues to function as a site of lexical innovation in current discourse. Accordingly, a term introduced several years ago may still qualify as a neologism if it: •remains thematised in public or specialised discourse; •has not yet undergone full lexical integration; •or is still perceived as novel by speakers, media, or institutional actors. This dynamic perspective allows for the inclusion of both recent coinages and older forms whose innovative character has persisted or re-emerged, and which is, for answering different research questions worth to work on. It also provides criteria for the exclusion of units that have become fully lexicalised, conventionalised, or obsolete. This perspective is reflected in the “Neologism of the Week” regularly published on the ENEOLI website,12 which highlight neologisms of particular relevance, interest, or discursive salience, regardless of their date of origin. The classification and selection criteria adopted by ENEOLI acknowledge the complex life cycle of neologisms, and the role of discourse, perception, and social uptake in shaping lexical innovation. By capturing both formal and contextual dimensions of new word formation, the Neologisms Collection offers a rich empirical basis for: •diachronic analysis of lexical trends; •contrastive studies across languages or domains; •the development of computational models of lexical innovation. At the same time, the description of neologisms entails a method for data-driven updates to NeoVoc. When network members describe the lexical innovation process undergone by a neologism, and its type, they are prompted with NeoVoc terms referencing concepts of the respective classes as controlled vocabulary items. In the event the description of the neologism at hand demands the inclusion of a new term, which may be special to the 12 See https://eneoli.eu. language of the described term, to its domain, or to the working language of the annotator, such additional term can be added to NeoVoc, which lives on the same platform. By linking a word sense of that term lexical entry to an existing or a new concept entry, the new term is integrated in the NeoVoc graph structure, without necessarily needing equivalent terms in other language; the new concept will be then declared, for example, a subclass of a broader concept, which may correspond to terms in other languages. 5. Data model summary The data model implemented for the ENEOLI Wikibase is designed to enable semantic interoperability both within the platform — across NeoCorpus, NeoVoc, and the Neologisms Collection — and with external knowledge graphs (see Fig. 1). The figure exemplifies the semantic integration at the core of the ENEOLI Wikibase. The three components are interconnected through typed relations that enable the construction of a shared knowledge graph. For example, the neologism pinkwashing (English), is classified as compound, which is linked from the lexical innovation process concept compounding, using a property called neologism type. The neologism pingkwashing (Spanish), in turn, is classified as borrowing from its homograph pendant in English. Fig. 1 employs different colours to distinguish between different types of entities. Bibliographic items are represented in light red, signalling their role as sources of attestation. In blue, lexical entries from NeoVoc (dark blue) and from the Neologisms Collection (light blue), while their corresponding senses are marked in yellow to highlight the semantic layer between lexical and conceptual dimension. Concepts are displayed in green — general concepts in light green, and NeoVoc metaterminological concepts in darker green. As explained in section 3, these are hierarchically structured through subclass relations. Persons, such as authors, appear in white, indicating their function as agents linked to publications. Wikibase entities can be aligned to external knowledge graphs such as Wikidata; this is true for bibliographical records, for parts of these records (e.g., journals), for items representing abstract concepts, for lexemes, and even for lexeme senses. While all data remains integrated in a common graph structure, SPARQL queries provided to users on text pages in ENEOLI Wikibase13 offer different partial views of the content, either visualizing one of the three contained resources independently, or showing relations between one and the other, either showing terms and glosses in the user working language, or lists of translation equivalents. To be able to enrich the different views of the data with content retrieved from Wikidata, some classes of ENEOLI Wikibase entries are systematically annotated with their identifier on that platform. Where references to other sources makes sense, we also add those. The alignments include: •linking author entities and bibliographic records in NeoCorpus to Wikidata identifiers, thus connecting ENEOLI bibliographic data with broader author profiles, publication histories, and affiliations, and enabling contributions to the Wikidata bibliographical data collection. 13 User documentation and queries are accessible through the main page at https://wikibase.cloud. Figure 1: Simplified graphical model of relations between entities in ENEOLI Wikibase •enriching NeoVoc lexical and concept entries with links to terminological databases or semantic networks; •linking NeoVoc and neologism entries to Wikidata lexemes; •aligning neologism entries with definitions or citations in external resources such as dictionary or language corpus portals, and documents describing the lexical item or the denoted concept (e.g.„ online encyclopaedia articles). The Wikibase query service allows users to trace the emergence and typology of neologisms across sources and languages, involving NeoVoc concepts in the query. In addition, ENEOLI Wikibase content remains interoperable for being involved in database queries from beyond the platform, through its open SPARQL endpoint.