Musicology as data science and heritage science

Cost Action CA 21161 DOI: 10.5281/zenodo.17540237 A New Ecosystem of Early Music Studies 5 MUSICOLOGY AS DATA SCIENCE AND HERITAGE SCIENCE This publication is based upon work from COST Action Earlymuse CA21161, supported by COST (European Cooperation in Science and Technology)”. COST (European Cooperation in Science and Technology) is a funding agency for research and innovation networks. Our Actions help connect research initiatives across Europe and enable scientists to grow their ideas by sharing them with their peers. This boots their research, career and innovation. Early Muse CESR-RicercarLab ISBN: 979-10-981515-07 Edited by Philippe Vendrix & Rebekah Ahrendt How to quote: Philippe Vendrix and Rebekah Ahrendt, eds, et al., Musicology as Data Science and Heritage Science, Tours, CESR, 2025. 2 – COST Action CA 21161, Earlymuse Action leaders: Philippe Vendrix (Chair), Rebekah Ahrendt (Vice Chair), Judith Haug (Grants Awarding Coordinator) WG leaders: Aleksandra Pister (WG Education), Grzegorz Joachimiak (WG Sources), Andrea Puentes-Blanco (WG Publications), Pedro Sousa Silva & Kalina Tomova (WG Performances), Brianne Dolce (WG Policies) Contributors: Rebekah Ahrendt, Ailin Arjmand, Marek Bebak, Angela Bellia, Niels Berentsen, Vincenzo Borghetti, Augustin Braud, Camilla Cavicchi, Jeremy Coleman, Emilie Corswarem, Brianne Dolce, Janine Droese, Mehmet Ugur Ekinci, Richard Freedman, Klemen Grabnar, Altea Guevara Alonso, Sonja Hamhuis, Judith I. Haug, Ivana Jelac a, Grzegorz Joachimiak, Maciej Jochymczyk, Lyra Kastrati, Metoda Kokole, Olivier Lartillot, Eduard Lazorík, Caro Lesemann-Elliott, Christophe Levaux, David Lewis, Magdalena Marija Measic, David Meredith, Fabian Moss, Marten Noorduin, Ana Oliveira, Angel Manuel Olmos Saez, Kimberly M. Parke, Vicente Parrilla Lopez, Anne Piejus, Guntars Pranis, Andrea Puentes Blanco, Laurent Pugin, Christophe Riedo, Emilio Ros-Fábregas, Gabriele Rossi Rognoni, Sonia Rzepka, Mark Saccomano, Thomas Schmidt, Laurie Stras, Hana Studenic ová, Martha E. M. Thomae Elias, Philippe Vendrix, Eva Veselovská, Vilena Vrbanic, Rebecca Wolf, Emily Worthington, Sonia Wronkowska Abstract: This collective guide explores the transformative intersections of musicology with data science and heritage science in the European context. It positions musicological research as both a model for and contributor to broader cultural heritage, analyzing its tangible and intangible dimensions. The report reviews advances in research infrastructures, including the European Open Science Cloud, ECHOES, and E-RIHS, highlighting challenges and opportunities for integrating music into transnational heritage frameworks. It surveys the complexity of music data — notation, audio, metadata, and visual sources — and emphasizes standardised methods for cataloguing and preservation. The text critically assesses legal, policy, and technical barriers to digitising and reusing musical materials, advocating for legal reforms and open science principles. New publication paradigms are proposed, leveraging digital tools to expand access and impact. The role of musicologists in public engagement, education, and fostering research communities is underscored as crucial for sustainability. Best practices and collaborative models link musicological data to broader digital humanities and preservation strategies. Ultimately, the guide calls for active participation in European policy and research networks to enhance musicology’s role in shaping and disseminating cultural heritage for future generations. – COST Action CA 21161, Earlymuse 3 Table of Contents Introduction 5 6 1. Infrastructures 5 9 1.1. European Open Science Cloud: EOSC > 10 1.2. European Cloud for Cultural Heritage: ECCCH-ECHOES > 11 1.3. European Research Infrastructure for Heritage Science: E-RIHS > 14 1.4. Europeana Pro and the Common European Data Space for Heritage > 15 1.5. Digital Research Infrastructure for the Arts and Humanities (DARIAH) > 17 1.6. Open Scholarly Communication in the European Research Area for Social Sciences and Humanities (OPERAS) > 18 1.7. Cultural and Creative Sectors/Industries (CCS/CCI) > 19 1.7.1. The role of musicology in industrial development > 20 1.7.2. The benefits > 20 1.7.3. Types of actions to be taken with the cultural and creative industries > 21 1.8. National and transnational initiatives > 22 2. Materiality: Data for Musicology 5 25 2.1. Recognising data for musicology > 25 2.2. Sources of data > 26 2.3. Types of data > 31 2.4. Structuring music data > 32 2.4.1. Metadata for musicology > 33 2.4.2. Encoding music scores > 38 2.5. Context of data management > 40 2.5.1. Collection and creation > 41 2.5.2. Usage > 43 2.6. Resilience > 46 2.6.1. Strategies for Protection and Digitisation > 48 3. Praxis 5 52 3.1. Organising > 52 3.2. Annotating > 53 3.3. Enriching > 57 3.4. Transforming > 59 – COST Action CA 21161, Earlymuse 4 4. Policy and Legislation 5 61 4.1. Introducing the legal and policy landscape > 62 4.2. Current challenges for musicology > 65 4.3. Fragmentation and lack of standards > 68 4.4. (How to navigate) challenges you might encounter > 69 4.4.1. Access: what can be digitised and made available? > 70 4.4.2. Use: what can be analysed with digital tools? > 70 4.4.3. Circulation: what can be shared across borders? > 71 4.4.4. Navigating the challenges: checklist for research projects > 72 4.5. Opportunities (and ways to get involved) > 75 4.5.1. Emerging infrastructures & opportunities > 75 4.5.2. Pathways for individual engagement > 77 4.5.3. Pathways for collective influence & engagement > 78 5. Outcomes 5 79 5.1. Education, Learning and Professional Development > 79 5.2. Linking and accessibility > 82 5.3. Reading and Writing > 83 6. New models of publication 5 87 6.1. Limitations of traditional publications > 88 6.1.1. Physical, linguistic, and economic limitations > 89 6.1.2. Notation as limitation > 91 6.1.3. Research environment as limitation > 94 6.1.4. Limitations in narrative structure > 94 6.2. Leveraging technological advancements for a new model of musicological publication > 97 6.2.1. Delivery and discovery > 98 6.2.2. Sustainability and Experimentation > 101 6.3. Addressing inequalities through new publication models > 102 6.3.1. The landscape of musicological publishing: existing inequalities > 102 6.3.2. Arguments for developing a new model > 103 6.3.3. Steps towards a tailor-made solution > 106 6.3.4. The European Accessibility Act (2019) > 106 6.3.5. Open Science and peer-review systems > 108 – COST Action CA 21161, Earlymuse 5 6.4. An economic model > 110 6.5. Long-term preservation > 113 6.6. Building public engagement into publication platforms > 115 7. Developing Communities 5 116 7.1. Building people > 117 7.2. Building institutions > 118 7.3. Building community > 119 7.4. Building Capacity > 121 8. Enhancing public engagement and impact measurement 5 124 8.1. Musicology in the public space > 125 8.2. Participation and Collaboration in Public Engagement > 127 8.3. Research Policy and Public Engagement > 128 8.4. Considering impact strategies in musicological research projects > 130 9. Communication and collaboration 5 135 – COST Action CA 21161, Earlymuse 6 Introduction The European Commission considers the music sector vital to safeguarding Europe’s cultural diversity and to strengthening its competitiveness. Music is the universal language of humankind and holds a unique, creative and cohesive power for both society and the individual. It also has great economic importance. The music sector is the third largest employer within the cultural and creative industries in the EU. Based on small and medium businesses, the European music sector is very dynamic and innovative, and it has an important growth potential.1 Cultural heritage encompasses both the tangible and intangible, and is a vital component of Europe’s cultural landscape and a key term for many funding agencies. Musical heritage, including both its material and sonic aspects, is an essential part of European cultural objectives. As both tangible and intangible heritage, music and its related products and systems of knowledge — most notably, musicology — can serve as both test cases and models for other branches of heritage. Yet, music remains underrepresented in both policy and heritage science infrastructures. Musicology and musical heritage are deeply interconnected. Musicologists not only study the history, theory, and cultural contexts of music, but they also actively shape how musical heritage is preserved, understood, and experienced today. Their work extends far beyond academic research: it informs and constructs editions of musical works, performances, recordings, audiovisual productions, heritage site programming, and public education initiatives, providing essential historical, analytical, and interpretive context. The work of musicologists has also enabled large-scale data projects on musical corpora and created innovative digital tools for the optical recognition of musical notation and analysis of musical structures, among others. Musicologists are involved in every step of these processes, often through publicly-funded research that ensures Europe’s rich musical past remains part of its present and future. But how can musical scholarship fit into ‘cultural heritage’ as it is currently used? The tools we employ for musicology shape our understanding of the past, and many musicologists continue to engage directly with materials from historical manuscripts, early prints, or physical archives, 1 ‘Music,’ European Commission. – COST Action CA 21161, Earlymuse 7 employing traditional critical practices. These approaches cultivate a profound sense of connection and attention. Yet, as our musicological heritage grows increasingly vulnerable with the passing of time, new approaches and tools are emerging that allow us to ask new questions and visualise new answers: for instance, we can preserve musical archives while simultaneously making them available to a wider audience through facilitating performances or generating data for analysis. How do we ‘make’ musicology into something that not only fits within but also contributes actively to cultural heritage? It requires a change of perspective, from that of a lone scholar working in a sharply defined field to a vibrant community which transforms scholarship into a living practice which attends to the past while looking toward the future. Musicology has a long and strong record of developing skills and (digital) tools by which to study music. The field is also aware of methodologies developed by other sciences and is able to use them. The very versatility of musicology — a broad field of science that encompasses the history, theory, and practice of music across time and geographical space — means that it has the strength and flexibility to model approaches to cultural heritage for other fields. Technologies of writing have always shaped the relationship between creators and users of text and media. Music is no exception. Throughout history, from early Western musical notation through the printing press, audio recording, and into today’s digital realm, emerging transcription technologies have provided mechanisms to regulate music’s impact and function, while simultaneously establishing it as a form of intellectual property. These innovations each represented the ‘new media’ of their era, creating fresh opportunities for interaction among composers, performers, and audiences centered on musical concepts. Developing musicology for the digital era demands careful consideration of musical texts in their full complexity, spanning from primary sources to digital displays, and encompassing written documents through to recorded interpretations. Beyond this, we must embrace novel approaches to conceptualizing music as data (via stratified, coded scores) alongside data concerning music (the extensively layered metadata contextualizing music within its creation and reception history). We want to think of our Musical Heritage as Data. But we need to think of Musicological Data as Heritage, too. * * * – COST Action CA 21161, Earlymuse 8 This document aims to accompany any musicologist in their desire to undertake research that positions the musical object as data science or heritage science. It brings together diverse resources, recommendations, suggestions for further exploration, and guidance. It attempts to account for the changing context in which musicology evolves, from its essential links with musical practice to its role in the cultural production chain. The collective text that follows claims nothing more than to serve as a guide to any musicological researcher wishing to position their work in relation to the complex fields that are data science and heritage science.2 The reader will therefore encounter a journey that moves from a description of European frameworks to the legal constraints that govern scientific production and artistic creation. The consequences of working with musicological data are felt in the way results are presented via new modes of publication, but also in how researchers experience their daily activities. This gives rise to new community spaces around contributors, the institutions in which they conduct their projects, and end users. The combination of these communities reconfigures the research space in which we operate. And this combination invites us to rethink the way in which a historical discipline, musicology, expresses itself in public spaces. The reader should not expect to learn recipes, because there are none. Nor should they hope to read an exposition of a consolidated methodology assuming a firmly structured epistemological perspective. The innovations that punctuate our daily lives require constant questioning. We acknowledge that it is impossible to claim exhaustiveness, impossible to escape the risk of obsolescence, impossible to cover in detail each of the challenges, each of the concepts, each of the technological innovations, but we hope that this guide will encourage readers to find their own paths toward the musicological future. 2 The themes discussed in this document were addressed during meetings and a collective Short Term Scientific Mission in 2023, 2024, and 2025. Meetings took place in Lisbon in March 2023 hosted by João Pedro d’Alvarenga, in Malaga in September 2024 hosted by María José de la Torre Molina, in Zagreb in February 2024 hosted by Ivan Curkovic , in Wrocław in May 2024 hosted by Grzegorz Joachimiak, in Ljubljana in March 2025 hosted by Metoda Kokole, in Bratislava in March 2025 hosted by Eva Veselovská, and in Brussels in September 2025 hosted by Christophe Levaux and Marie Cornaz. – COST Action CA 21161, Earlymuse 9 1. Infrastructures In recent years, there have been a number of initiatives to network European research and knowledge around data science and heritage science. The way we approach and deal with cultural heritage is evolving rapidly thanks to digital technologies, and it is important to bring these developments into the next digital decades. The strategic approaches of the European Commission and the Member States of the European Union to the digitisation of cultural heritage can be identified in a number of national and supranational documents, platforms, frameworks, and EU projects (ECHOES/ECCCH, INCEPTION, DARIAH-EU, SCRIBEMUS, MusicAIRE, eKultura, etc.). There are two primary challenges to the integration of musicology in such infrastructures and projects. The first is a lack of awareness amongst musicologists of the need to respond to these projects even when music is not specifically mentioned, or perhaps insecurity about what projects might best fit. We will address this challenge further below. A more pressing challenge is the need to address basic legal and policy issues and to begin to develop the standards and tools that will enable music heritage and musicology to fully participate in such initiatives. Participation and incorporation of musicology into existing or developing infrastructures can be facilitated by data sharing and data interoperability. Here we can distinguish two main types of data resources that need to be attended to. Firstly, we have numerous databases and resources that musicologists have been developing and populating through decades of research. Integration of these pre-existing resources will be possible through promoting the adoption of linked open data (LOD) models and technologies. Many actions have already been undertaken, including the international LinkedMusic project, but there is still much to be done. Here we can distinguish two directions. On the one hand, fostering the use of reference data models that have been developed and adopted widely by cultural heritage institutions, such as CIDOC-CRM or IFLA LRM (see below). On the other hand, database integration into larger datasets will require music domain-specific authority resources to be put in place and curated. For instance, authorities for music personal names, musical instruments or musical works are all resources to which musicological research can contribute. The Répertoire International des Sources Musicales (RISM) project has assumed a central role in authority resources, particularly through its international and cross-border structure. Scoping and sharing such authority – COST Action CA 21161, Earlymuse 16 develop and openly share software tools, datasets, and ontologies designed to foster new research and spur innovation in both musicology and digital humanities. Ongoing professional development is also key. Musicologists can contribute to webinars, conferences, and workshops organised by Europeana Pro, sharing expertise around data workflows, Linked Open Data, AI, and 3D technologies in cultural heritage. Collaboration on designing standardised metadata schemas and interoperability standards is essential to improving data sharing across diverse music collections, ensuring that resources are compatible, discoverable, and useful to a global community. Ultimately, these initiatives contribute to the Common European Data Space for Heritage. This is a major European Union initiative designed to accelerate the digital transformation of the cultural heritage sector by providing a secure, interoperable infrastructure for sharing, accessing, and reusing digitised cultural heritage data throughout Europe. The project aims to digitise monuments, sites, objects, and artefacts, making high-quality 2D and 3D data widely available to safeguard at-risk heritage and support reuse in areas like education, tourism, and the creative industries. It is a flagship action built on the Europeana platform, which already offers access to millions of digitised cultural assets. Member States are encouraged to digitise all monuments and sites at risk of degradation and at least half of the most-visited locations by 2030. GRAPHIA aims to significantly improve data quality, actionability, interoperability, and visualization in SSH by converting unstructured data into actionable insights with AI, facilitating advanced analysis and discovery of complex patterns and connections within SSH fields. – COST Action CA 21161, Earlymuse 17 In the Data Space, music occupies a significant place, although its representation is still developing compared to images or texts. Music is part of the vast body of digital heritage resources that are aggregated and made accessible through the Data Space, mainly via initiatives such as Europeana Sounds and specialised portals like Polifonia or the Open Music Observatory. Europeana, which serves as the backbone of the infrastructure, currently offers more than 59 million digital objects, including millions of recordings and sound archives, which can be consulted and reused in educational, scientific, or creative projects. The Open Music Observatory embodies a decentralised approach to aggregation, enabling federated sharing and the exploration of musical metadata according to European data space standards, while also encouraging interdisciplinary connections with other heritage domains. 1.5. Digital Research Infrastructure for the Arts and Humanities (DARIAH) DARIAH is the Digital Research Infrastructure for the Arts and Humanities, a European consortium that enhances and supports digitally-enabled research and teaching across the arts and humanities by providing tools, services, data, and expertise for researchers and educators. DARIAH is structured as a European Research Infrastructure Consortium (ERIC). Its primary aim is to foster collaboration, preserve and disseminate research outputs, and develop technologically advanced infrastructure for research practices in the digital arts and humanities. Through its national branches, DARIAH has funded a wide range of music-related projects, spanning artistic practice, music technology, and digital heritage. Among the many initiatives supported by DARIAH, it is worth mentioning MusicPUT. MusicPUT is a specialised digital musicology initiative operating within DARIAH-PL, which is the largest humanities consortium in Poland. The team based at the Institute of Computer Science at Poznan University of Technology focuses on creating innovative computational tools for music research and analysis. The – COST Action CA 21161, Earlymuse 18 MusicPUT platform represents a significant contribution to digital musicology by providing specialised tools that bridge computer science and music research. Their products serve various musicological needs: • ScoreScribe: Automatic transcription from audio to symbolic formats (MIDI, MusicXML, MEI, ABC), particularly valuable for ethnomusicologists working with field recordings. • OMRAT: Optical music recognition for automatic annotation of sheet music, enabling content extraction and indexing for music libraries. • Timbra: Timbre analysis and visualization for specialists studying sound characteristics. • MEIConvert: Batch conversion between symbolic music formats. The platform specifically addresses musicologists, music librarians, ethnomusicologists, sound engineers, musicians, and educators who lack programming expertise, democratising access to advanced computational musicology tools. The platform’s integration within DARIAH-PL ensures sustainability and connection to the broader European digital humanities infrastructure, facilitating international collaboration and resource sharing. The MusicPUT project exemplifies how national digital humanities consortiums can develop specialised disciplinary tools while contributing to the larger European research infrastructure ecosystem. 1.6. Open Scholarly Communication in the European Research Area for Social Sciences and Humanities (OPERAS) Open Scholarly Communication in the European Research Area for Social Sciences and Humanities (OPERAS) is a research infrastructure specifically geared toward open scholarly communication in – COST Action CA 21161, Earlymuse 19 SSH fields in the European Research Area. Its mission is to coordinate and federate resources across Europe in order to facilitate accessing, creation, edition, dissemination, and validation of SSH research outputs. It aims to operate efficiently and openly, without barriers for researchers, students, or the public. • OPERAS is designed as a distributed infrastructure, involving a broad network of partners and nodes throughout Europe. • Key services developed under OPERAS include the Directory of Open Access Books (DOAB), authentication and authorization solutions for research resources, and technical as well as social innovation labs to support new communication practices and governance models in SSH. OPERAS intends to overcome the fragmentation and underfunding in the SSH scholarly communication landscape—especially challenges like multilingualism, diverse (open) publication formats, and the need for specialised scholarly outputs (monographs, critical editions, etc.). OPERAS provides a service portal and a set of innovative tools to facilitate open access publishing, discoverability, certification, and research metrics in SSH. The infrastructure is supported by EU-funded projects such as OPERAS-D (Design), OPERAS-P (Preparation), and OPERAS-PLUS, which further its development. 1.7. Cultural and Creative Sectors/Industries (CCS/CCI) European heritage and industrial strategies identify music as a key component. As part of the Cultural and Creative Sectors, music-related initiatives are supported by Creative Europe and by the specificallytargeted music sector framework Music Moves Europe. Music is also an unmissable industry within the Cultural and Creative Industries, one that has received a great deal of policy attention from the European Commission in recent years. Responding to calls for proposals linked to the CCS or CCI strategies requires a threefold reflection: firstly, on the role of musicology in industrial development; secondly, on the benefits that each party (research and industry) can expect from it; and thirdly, on the actions to be taken.5 5 See ‘How do UK research funders foster completion, collaboration, and impact (CCI) in international development research calls and programmes?’ – COST Action CA 21161, Earlymuse 20 1.7.1. The role of musicology in industrial development Tapping heritage for modern audiences: The Cultural and Creative Industries’ and society’s understanding of European musical heritage remains limited and often tied to outdated, rigid ideas of European identity. As a consequence, the substantial potential of this heritage that could directly inform creative and cultural products remains untapped. Engaging contemporary, diverse audiences in innovative ways and addressing their concerns and lived experiences while fostering a sense of cohesion and identity is a task musicologists are equipped for. Seizing opportunities for sustainable integration: The specialist expertise of musicologists and the evidence base their work could provide are often not fully integrated into the Cultural and Creative Industries sector. This leads to missed opportunities in creating a sustainable knowledge economy where research and musical understanding are accessible to benefit a wide range of cultural and creative industries, as well as tourism, built heritage and territorial development. Realising potential in cultural development: The European Cultural and Creative industries are a major economic output from the region, but there is significant unrealised potential for musicologists to contribute to evidence-based and data-driven strategies in cultural development. By enabling mutual benefits for musicologists and the Creative and Cultural industries, there is scope to drive innovation in cultural heritage and development that are grounded in recognised specialist world-leading scholarship. 1.7.2. The benefits Empowering Cultural and Creative Industries with accessible knowledge: The Cultural and Creative Industries will be able to quickly and efficiently tap into a range of knowledge, data and resources that will improve their capacity to respond to the needs of local and international societies, creating innovative and relevant cultural products and services. Fostering sustainable careers for musicians: Musicians and other agents in the musical and cultural ecosystem will be able to develop more stable and sustainable careers and be empowered by wider access to heritage resources to shape their own creativity. – COST Action CA 21161, Earlymuse 21 Linking heritage and Cultural and Creative Industries: Heritage organisations will develop stronger connections with the cultural and creative industries, increasing their capacity for innovation and allowing heritage to be reimagined in ways that resonate with contemporary audiences, creating new opportunities for research, preservation and commercial development within the creative sector. Promoting European development through music: European territories (e.g. towns, cities, regions) will be able to strengthen the links between Europe’s musical heritage and their own role in boosting and revitalising territories —for example, by offering more sustainable and resilient tourism products/ services. Enhancing access and well-being through culture: The users of the cultural and creative industries, as well as the broader society, will have access to an increased, more relevant offer, leading to increased wellbeing, greater awareness of the diversity of European identities and greater social cohesion. 1.7.3. Types of actions to be taken with the cultural and creative industries Clearly define ‘musicologist’ as an official job category: By recognising musicologists as specialised professionals (e.g., within OECD categories), policymakers can more easily quantify the contribution of musicology to evidence-based policymaking. Musicologists provide essential data-driven insights that can assist policymakers in understanding the specific impacts and benefits of musical research on the wider cultural and creative industries. This recognition ultimately supports the development of more effective policies for the sector. Support nomination of musicologists in advisory boards: When creating policies, consider including musical heritage and its manifestations as vital components in promoting diversity and territorial development. Musicologists bring essential expertise and insights that can inform decision-making, ensuring that cultural policies are rooted in a comprehensive understanding of musical traditions and manifestations, and their impact on communities. Develop a user-friendly data gateway: To enhance the music sector and public engagement, a userfriendly gateway providing access to high-quality, open-access musicological data is essential. This gateway would serve as a comprehensive resource, sparking innovation and enabling the use of valuable music data in new ways to future-proof information within a European context. The creation – COST Action CA 21161, Earlymuse 22 of an interdisciplinary network of musicologists, musicians, web designers, and developers would be fundamental. This team would collaborate to ensure the gateway’s technical and artistic quality. Create micro-grants for knowledge exchange: Establish micro-grants to support and incubate knowledge exchange in scholarly-creative partnerships. These initiatives could contribute to a repository of best practices and may also include an innovation/technology priority, such as proposing different ways to integrate AI into projects related to making musicological data available. 1.8. National and transnational initiatives Building on and contributing to the models and data of significant national and transnational initiatives will help move our work into the musicological future. Here we would like to highlight a few representative examples regarding data science and cultural heritage. In Germany, the National Research Data Infrastructure (NFDI) is the umbrella structure of sub-discipline consortia for data preservation. The branch dealing with heritage is NFDI4Culture - Consortium for Research Data on Material and Immaterial Cultural Heritage, coordinated by the Mainz Academy of Sciences and Literature. Musicology is an important community within this consortium, and a recent report shows music is playing a central role in this context. The purpose of the consortium is to develop services, standards, and training for managing, preserving, and sharing both material and immaterial cultural assets, with musicology as a central community. For music researchers, it provides infrastructures for complex data types (scores, audio, video, 3D models), promotes FAIR principles, supports standards like MEI, and offers legal guidance on copyright and performer rights. Major musicological initiatives co-ordinated and supported by NFDI4Culture include the Centre for Digital Music Documentation (CDMD) which itself acts as a clearinghouse for digital initiatives within the musicological projects funded by the German Academies of Sciences. In the United Kingdom there is no official national research data infrastructure, but JISC (Joint Information Systems Committee), jointly sponsored by the government and participating institutions, acts as a de facto national provider of digital research infrastructure, including eduroam, a consolidated national library catalogue (formerly COPAC) and JANET as the national research data network for research and education. – COST Action CA 21161, Earlymuse 23 In France, Huma-Num is a national research infrastructure dedicated to the humanities and social sciences (HSS) and to digital humanities. It supports researchers in the management, storage, sharing, and preservation of digital data related to their HSS projects. Huma-Num provides digital tools and services throughout the entire research data lifecycle: secure storage (ShareDocs, Huma-Num Box), long-term archiving, dissemination (virtual machines, Nakala), indexing (Isidore), and interoperability. It promotes the principles of open science and FAIR data management. Within Huma-Num, MUSICA2 is a consortium specialised in digital musicology that brings together researchers and engineers around issues related to the management and modeling of musical and musicological data. The consortium’s main objective is to open up the discipline to interoperable and shareable approaches, while developing tools and standards suited to the specificities of musical sources (manuscripts, scores, archives, corpora, etc.). MUSICA2 has developed a cross-disciplinary search tool, MusicaSearch, and publishes best practice guidelines. Spotlight: MUSICASEARCH The Musica2 consortium has developed MusicaSearch, a search engine for digital musicology, to give greater visibility to digital projects and corpora scattered across France and abroad, attached to multiple, disconnected research units and organizations. To this end, MusicaSearch facilitates cross-searching in various Musica2 partner corpora, by displaying the results of a query in the form of an interactive list. MusicaSearch also invites you to create links between projects, by building active gateways from the data itself (partner site widget). MusicaSearch queries the data (people, ensembles, works, events) displayed on the various partner sites, or even the music data on generalist sites, offering an access point (API, OAI-PMH, Sparql endpoint). MusicaSearch aims to provide users - whether students, researchers, teachers, music lovers, musicians or the simply curious - with a centralised search tool for consulting data available in the field of digital musicology. – COST Action CA 21161, Earlymuse 24 The PEPR ICCARE is a priority research program and facility launched in June 2024, dedicated to cultural and creative industries (CCI) in France. ICCARE aims to support the CCI sector (visual arts, audiovisual, music, performing arts, design, architecture, heritage, video games, publishing, arts and fashion crafts, etc.) in its transformation in response to digital, ecological, and economic challenges. It forms the research component of the national CCI acceleration strategy, aligned with France 2030. Music holds a central place in ICCARE, both in strategic orientations and in concrete initiatives supported by the program. The Cultural and Creative Industries sector explicitly includes music, along with audiovisual, visual arts, publishing, and performing arts. ICCARE aims to bring together musicologists, music professionals, technicians, creators, and institutions for reflection and co-construction on contemporary issues in music. In Sweden, the Royal Swedish Academy of Music maintains an extensive website on Swedish musical heritage (Levande musikarv — living musical heritage), containing a database, downloadable sheet music, and scholarly publications with the aim of safeguarding, highlighting and popularising Swedish music heritage (both art and folk music). Similarly, The Fryderyk Chopin Institute in Warszawa, Poland, developed (with the support of EU funds) Polish Music Heritage and Polish Scores portals with the aim of making available metadata, scans, and digital transcriptions of musical sources held by several institutions. On a transnational level, Saving Ukrainian Cultural Heritage Online (SUCHO) is a multi-institution initiative to safeguard the digital cultural heritage of Ukraine. – COST Action CA 21161, Earlymuse 25 Spotlight: Saving Ukrainian Cultural Heritage Online (SUCHO) SUCHO holds valuable lessons for cultural heritage preservation efforts, particularly in highlighting how digital resources are also at risk of damage or destruction. The project relied on word-ofmouth through the digital humanities community to identify Ukrainian collections accessible online. Volunteers suggested collections to target for preservation, then project facilitators evaluated those collections according to their cultural importance and risk level. The shock and enormity of war in Europe managed to mobilise a large number of volunteers in a very short period. As an independent collective, SUCHO was able to set up its own servers and get started archiving right away—a much quicker process than going through institutional channels. Once the process was underway, the archiving process itself revealed additional candidates for archiving, including new domains to explore. This case illustrates the need to begin conservation conversations now, as well as the role to be played by informal, self-sustaining networks in identifying and archiving materials. 2. Materiality: Data for Musicology Unlike many kinds of data, music data is never ‘raw’. Sound being ephemeral, all music data we engage with is an encoding of a temporal physical experience in a defined space, whether it be notated music, a song text, a musical instrument, or choreographic instructions. Therefore, all music data is always inherently mediated, by humans or machines. This section is therefore concerned with when and how musical heritage becomes data. Terminology in this document is generally in line with the DIKW pyramid, explaining relations between data, information, knowledge and wisdom. 2.1. Recognising data for musicology Musical heritage sources represent the entirety of human involvement with making organised sound - sources that can be discovered, explored and made accessible through digital versions, innovative analytical methods and techniques. They are preserved either in tangible, physical forms (written notation, musical instruments, information about music in archival documents, literature, visual arts – COST Action CA 21161, Earlymuse 32 of texts from letters to concert programs to treatises or songbooks. There is cartographic data in research on music migrants and migration or on tracing musical traditions. There is also data created by librarians, archivists or museum professionals, in the form of handwritten inventories, card catalogues, printed catalogues, digital catalogues. Following established data-related practices and expertise in the humanities and heritage science, in an effort to organise the described abundance and complexity, the following broad types were identified in data for musicology: • Materiality • Notation (notes, symbols, words) • Audio data • Visual data • Personal data • Paratexts • Authority data • Access data • Bibliographical data These categories are not inherently divisible between ‘potential’ or ‘actual’ data. For example, a RISM record for a source provides both potential data (e.g., which sources feature a certain song setting, to be discovered) and actual data that can be linked with other actual data points (e.g., how a certain song setting is transmitted in several sources). By understanding the scope of music data as comprising information pertaining to broad themes, types or subdomains, we can recognise the fluidity of music data, with data points interacting with each other often on multiple levels at the same time. 2.4. Structuring music data Mapping and cataloguing sources is one of the fundamental tasks for both musicology and heritage science. Librarians, musicologists, and researchers have developed various models and standards for documenting sources, which ensure consistency, interoperability, and accessibility. Used for both – COST Action CA 21161, Earlymuse 33 tangible and intangible musical sources, these standards ensure the preservation and accessibility of valuable musical artefacts. 2.4.1. Metadata for musicology Since the majority of music sources (written music sources, music-related textual materials, recordings) fall under the category of library collections, the library perspective of organising data will be outlined at the first place, with elaboration on museums, archives and galleries later in this subchapter. The Library Reference Model (LRM) propagated by the International Federation of Library Associations and Institutions (IFLA) replaced the earlier Functional Requirements for Bibliographic Records (FRBR) model. LRM alongside Machine-Readable Cataloguing (MARC) and Bibliographic Framework (BIBFRAME) standards provides structured methodologies for organising, cataloguing, and digitally preserving musical sources, facilitating their long-term accessibility and scholarly use. The LRM model offers a conceptual structure for bibliographic data. It enables a nuanced understanding of the relationships between different manifestations of a musical work, which is crucial for the study and preservation of historical musical documents. MARC (Machine-Readable Cataloguing) and BIBFRAME (Bibliographic Framework) are metadata standards used for documenting and preserving musical heritage. While MARC has traditionally been used for cataloguing library holdings, BIBFRAME introduces a more flexible, linked-data approach, enhancing accessibility and interoperability in digital environments Despite its structured approach, MARC has limitations in handling complex historical relationships and linked data, prompting the shift towards BIBFRAME. BIBFRAME restructures bibliographic data into a linked-data model, offering improved capabilities for preserving and interconnecting musical heritage materials. Its three core classes align with LRM and historical musicology: Work, Instance (Manifestation in LRM) and Item. By leveraging linked data principles, BIBFRAME enhances the discoverability of musical sources across institutions, enabling researchers to access and compare materials from different collections seamlessly. LRM modelling of the relationships between Work, Expression, Manifestation and Item. – COST Action CA 21161, Earlymuse 34 The formats discussed above that are used in libraries are mainly applicable when describing tangible material sources. In the case of intangible sources, it may be more relevant to use standards developed by archivists, museum professionals or ethnographers. Museums, galleries, and archives often use different metadata standards suited to their needs and ensure proper documentation, retrieval, and interoperability of cultural heritage materials including intangible musical sources through the use of standardised dataframes and/ or thesaurii. Ontologies such as CIDOC CRM, the Conceptual Reference Model (CRM) created by the International Committee for Documentation (CIDOC) allow Linked Open Data and are favored for their innate interoperability There is currently no widely recognised metadata standard for musical sources that is more specific than the established standards for libraries outlined above; higher level standards such as the Dublin Core Metadata Initiative (DCMI) general metadata standards for describing digital resources are not always sufficient. Museums, unlike libraries, primarily use objectcentred metadata standards for cataloguing artifacts, including musical instruments, recordings, and audiovisual materials. Common metadata standards include: RISM Online API response in MARCXML for source 600033396. CIDOC CRM high level ontology for main concepts in cultural heritage. Version 7.1.3, 2024. Licence CC BY 4.0. – COST Action CA 21161, Earlymuse 35 • CDWA (Categories for the Description of Works of Art) — Used for describing artworks, including musical instruments. • IconClass - Classification system for the content of images. • LIDO (Lightweight Information Describing Objects) — A metadata format for exchanging museum data across institutions. One of the major contributors working with musical instruments and using the above formats is MIMO - Musical Instrument Museums Online. The consortium’s current goal is to become the single point of access to information about public collections of musical instruments for the whole world and is used by musicologists working as much on acoustic instruments as on electronic instruments. Galleries, especially those dealing with digital art and audio-visual materials, use metadata standards similar to museums: • VRA Core (Visual Resources Association Core) — Used for describing images and multimedia, including performance documentation. Archives manage both physical and digital records, often focusing on structured metadata for preservation and retrieval. • EAD (Encoded Archival Description) — XML-based standard used to describe archival collections, including sound and music materials. • ISAD(G) (General International Standard Archival Description) — A framework for structuring archival metadata. • ISAAR CPF (International Standard Archival Authority Record for Corporate Bodies, Persons, and Families) — Provides authority control for describing entities related to musical archives. • MODS (Metadata Object Description Schema) — XML format for describing digital collections in archives and libraries. • METS (Metadata Encoding and Transmission Standard) — Used to store descriptive, administrative, and structural metadata for digital objects in archivers and libraries. MIMO: musical instruments museum online, catalogue search example. – COST Action CA 21161, Earlymuse 36 Given that many museums and archives store intangible musical heritage, various specialised metadata standards for Audio-Visual and Musical Archives are used: • PBCore (Public Broadcasting Metadata Dictionary Core) — Developed for audio and video materials in broadcast archives. • EBUCore — provides a common set of metadata for describing radio & TV media assets. • AES57 (Audio Engineering Society Metadata Standard) — Describes audio recordings for professional and archival use. • IAML’s UNIMARC for Music — Used in specialised music archives for musical sources and sound recordings. One of the most important and most successful recent initiatives for fostering data access and data reusing within cultural heritage institutions is undoubtedly IIIF, the International Image Interoperability Framework. Its development has been a huge boost to data sharing and data re-use for libraries, museums and archives. In addition to that, its organization as a consortium with multiple contributing institutions is also a funding model that cannot be ignored. Of course, musicology is by far too small for considering setting up a specialised framework. Nonetheless, IIIF can serve as a model for making musicology data, and encoded music scores in particular, accessible and re-usable in similar ways. To some extent, the Music Encoding Initiative model (see below) already fulfils this purpose and it can certainly serve as the basis for such a model. However, it is still necessary to think through access protocols and develop reference implementations for it in the same way the IIIF community developed reference image viewer implementations. Along the same lines as IIIF, another model from which we can draw inspiration from is the Distributed Text Services (DTS). DTS is in a way a IIIF for text. It is an open standard that defines a web-based API for accessing, querying, and retrieving digital texts in a structured and interoperable way. It is particularly useful in the humanities, enabling scholars, researchers, and developers to work with texts from different repositories without being tied to specific platforms or formats. DTS allows for navigation, citation, and metadata retrieval while supporting dynamic text collections and multilingual Possibilities of data exchange and integration through the implementation of IIIF in various heritage institutions, source: https:// training.iiif.io/iiif-5-day-workshop/ . – COST Action CA 21161, Earlymuse 37 content. Its goal is to enhance accessibility, interoperability, and reusability of textual data across digital libraries and research projects. Anastasimatarion (Duke University, MS emsgk01019) displayed via IIIF manifest using Mirador. In addition to the development of IIIF, IIIF-AV, a sub-project community group, should also be followed. In this context, there is a huge potential for encoding music scores to become a new type of interaction layer for accessing and navigating sound archives and for integrating Music Information Retrieval6 (MIR) techniques into cultural institution portals. Finally, the development of text publishing frameworks such as the TEI Publisher should also serve as a potential model for advancing data publishing in musicology. TEI Publisher has made it significantly easier for researchers to create, share, and explore structured texts in a sustainable and interoperable way, and this is something that is still lacking for musicology. 6 Music information retrieval (MIR) is the interdisciplinary field focused on the extraction, analysis and retrieval of meaningful information from music data. – COST Action CA 21161, Earlymuse 38 Demonstrating best practices and project models is also essential in the field of promoting musical heritage. From this perspective, it is important to build on work that could be recognised as model solutions in future EU programs. These include, for example, the examples set by RISM, MusicLabs, Verovio, Music-focused Transkribus OMR projects, OMMR4all. By adopting standardised metadata structures (e.g., MEI or MusicXML), by improving digital tools such as Optical Music Recognition (OMR), and by investing in pilot projects focused on the preservation and dissemination of digital content, best practices can be developed and disseminated. 2.4.2. Encoding music scores The XML-based Music Encoding Initiative (MEI) has established itself as the most advanced score encoding framework for digital musicology research and beyond. MEI has been adopted by dozens of musicology research projects, infrastructures and corpora. The use of MEI ranges from simple scores in common western music notation (CMWN) to very rich critical or genetic music encoding. Beethovens Werkstatt project at the University of Paderborn and the Beethoven Haus in Bonn where MEI is being used for encoding the genetic process of Beethoven’s sketches. – COST Action CA 21161, Earlymuse 39 An important step to make for the MEI community will be clarification and consolidation, not primarily of the model itself, which is already extremely advanced and powerful, but of the encoding level and the encoded content. A recent publication by Kévin Roger summarises state-of-the art use cases and workflows in MEI.7 Further to this, music data can be formatted in a wide variety of ways, both digitally and physically. Formatting styles can be generally categorised as involving standards for formatting symbolic music notation, metadata, text, image, audio, and visual representation. Examples pertaining to these categories are provided below. MEI can carry a great quantity of metadata related to the musical sources. An initiative aiming for standardisation is currently being undertaken by the edirom project based at the University of Paderborn through their MEI-based MerMEId editor — it is being adopted by various digitallyinformed edition projects in Germany in particular, but it remains to be seen whether this will be widely accepted. The absence, for now, of a universally-accepted standard for music data means that some data pertaining to music-specific aspects risks being lost or captured in unsatisfactory ways. Other areas within the heritage sector have developed their own standards, beyond the general ones, which could in part serve as a model for musicology. Another recent initiative is the publication of a Guide for digital musical editions8 published by the french 7 Kévin Roger (2025), ‘The Music Encoding Initiative: Its Generation Workflows’. 8 Philippe Vendrix, Augustin Braud (2025), Guide de l’édition musicale numérique. MerMEId is a tool for editing, handling, and displaying metadata based on the XML schema of the Music Encoding Initiative. Symbolic music notation formatting standards MEI suited for critical editions Humdrum suited for statistical analysis EsAC (Essen Associative Code) suited for analysis and comparison of traditional music on Central Europe MusicXML suited for exchange between music editor software Plaine & Easie Code suited for easy manual encoding of music incipits – COST Action CA 21161, Earlymuse 40 Consortium Huma-Num Musica2. It provides a structured survey of methods and standards for the digital editing of notated music, emphasising interoperability, sustainability, and FAIR principles. The main topics include notation software and encoding formats (MusicXML, MEI, Humdrum), conversion and visualization processes, long-term archiving, and metadata integration to ensure traceability and scholarly citation. Situated in the context of European cultural policies and digital research infrastructures, the Guide combines technical guidance with institutional perspectives, offering a practical and research-oriented framework for digital music editing. Excerpt of a MusicXML score at measure 3 showing a quarter-rest followed by a quarter note F4 in voice 1 on staff 1. 2.5. Context of data management Data science approaches to musicology highlight the overlaps between historical musicology, systematic musicology, and music theory and analysis, further drawing into question the dichotomies previously constructed between subfields. Data science approaches enable interrogation of past tendencies in musicology to emphasise the uniqueness of a source or work. Data science can also – COST Action CA 21161, Earlymuse 41 provide reliable statistics about the background against which musicological claims can be made, so long as parameters are established to ensure accurate conclusions are drawn from the data analysis, as will be discussed in this section. 2.5.1. Collection and creation In order to help musicologists work out the parameters for data collection in a given project, it is useful to have broader parameters for the fundamentals of data collection, in part to avoid entrenching hierarchies of data types. Previous broad parameters for music data collection have often relied on assumed inherent hierarchies regarding data types, and have often obfuscated the role of historical agents and communities that have - and continue to be - marginalised in society (e.g., women, ethnic, religious and other minorities, queer people, etc.). Given the importance of all kinds of agents and communities to music history, such marginalisation results in a skewed picture. To avoid inherent hierarchies, we have proposed a set of questions which musicologists should consider while establishing the context for data creation in their studies. • What is the purpose for creating data (e.g., conservation, preservation, a specific study, facilitating future study)? • What are the ethical and practical implications of the data creation? • What historiographical approach does the study in question pursue? (e.g., book history, material history, religious history, organology, etc.) • What role does chronology play in the study? • What role does geography play in the study? • What are the practical restrictions facing the researcher(s) in creating data? Some examples of the context or motivation for data creation include: • Digital preservation of tangible or intangible heritage • Presentation and indexing • Enabling content search • Analysis and processing • Constructing information and knowledge – COST Action CA 21161, Earlymuse 48 Many scientific institutions have problematic profiles that are subject to frequent changes and different approaches (loss of identity of scientific disciplines, pursuit of ‘current’ and ‘popular’ topics in scientific fields, etc.). We emphasise that an important part of managing the risks we mention here is for experts in musical heritage–like musicologists–to become involved in developing preservation and digitisation strategies. 2.6.1. Strategies for Protection and Digitisation The importance of digitisation as a preservation strategy has been on the European agenda for some years now. Already in 2019, twenty-seven European countries signed a Declaration of cooperation on advancing digitisation of cultural heritage, and Europe has continued to develop extensive digital policies ever since. Musicology has historically been at the forefront of developing digital tools and projects that contribute to the preservation of musical heritage. The oldest initiative that aims to comprehensively document extant tangible musical sources (manuscripts, printed music editions, writings on music theory, and libretti) worldwide is the Répertoire International des Sources Musicales (RISM) - International Inventory of Musical Sources, founded in 1952 in Paris. Today, over 1.7 million records can be searched at no cost through the Catalogue and RISM Online. RISM records what exists and where it can be found,10 and has become one of the main universal and reference tools for musicologists to find and document musical sources. 10 Andrew Hankinson and Laurent Pugin (2024). ‘Navigating the RISM Data with RISM Online’. Incipit search in RISM Online provides access to detailed information about music sources held in thousands of libraries and archives in Europe and abroad. – COST Action CA 21161, Earlymuse 49 In recent years, RISM Online has begun to assimilate data from other important musicology projects, allowing for integrated searches. A notable recent addition to RISM searches is the Digital Image Archive of Medieval Music (DIAMM), which is a great example of a musicology-led project on cultural heritage preservation that has also become a major metadata resource. Nationally-based projects and initiatives, such as Books of Hispanic Polyphony or the Portuguese Early Music Database provide comprehensive information about early music sources in a given country, often going beyond the data recorded by RISM. They remain essential for their contributions to the scientific knowledge of musical sources and their preservation. In future, it would be desirable to link the valuable data of these projects in an interoperable system. Other resources have begun to incorporate data and digital twins of musical heritage resources. We might return to the example of Europeana, Europe’s digital platform for cultural heritage, which has made available over 50 million digital records from the collections of thousands of cultural institutions DIAMM catalogue record for A-la Inventar no. KK 5370–5373; black void mensural notation; shows shelfmark, measurements, identifiers, and authority links. https://www.diamm.ac.uk/sources/1834/#/. The Books of Hispanic Polyphony database (catalogue search example) is an open-access catalog with linked images and detailed information about Spanish polyphonic sources. – COST Action CA 21161, Earlymuse 50 across Europe, including sources of musical heritage. This data was collected by various participating organizations (like libraries and museums). However, the data is not always of high quality, due to the particular metadata standard that Europeana adopted.11 This standard does not fit musical objects and resources very well. We could say that, due to its multimodal nature, the digital twin of a musical source would need at least three items: an image of the source (IIIF standard comes to mind), an encoding of the musical information and a recording of its performance, the three of them including all necessary metadata. Europeana aggregator, results from various libraries to search query for a composer’s name. 11 Carlotta Capurro, Gertjan Plets, and Jaap Verheul (2023), ‘Digital heritage infrastructures as cultural policy instruments: Europeana and the enactment of European citizenship.’ – COST Action CA 21161, Earlymuse 51 In all such projects, we strongly hold that expert musicological researchers should be involved. We also would like to indicate that those involved in developing such resources be aware of the following global issues: • Metadata and Cataloguing. An important step is the local unification and global compatibility of standardised metadata and standard formats (e.g. Dublin Core, MARC, FRBR, BIBFRAME etc.) in order to speed up and facilitate the retrieval of materials and their content. • Digital Conversion and Archiving. The primary objective is to ensure efficient, long-term, and high-quality digitisation of tangible and intangible music sources. Storage needs to be implemented in multiple formats and backups to prevent data loss. • Online Accessibility and Open Archives. The creation of digital libraries and platforms to facilitate global access to musical sources at national or transnational level is a fundamental precondition for a comprehensive elaboration of musical cultural heritage. Cooperation and compatibility of systems with institutions such as UNESCO, national archives, libraries (inventories: RILM, RISM, CANTUS INDEX network databases, DIAMM etc.) is fundamental. • Community Involvement and Ethical Considerations. The processes assume the involvement of local/local communities (e.g. religious communities, private collections, etc.) in synergy with ensuring the ownership rights Most importantly, though, for any digitalisation project to be truly successful, it must result in data that is machine readable, interoperable, and reusable. That means it must be standardised. We expect greater use of artificial intelligence and machine learning applications for sound recognition and historical analysis with flexible encoding (supports both modern and historical notation systems), with metadata-rich encoding (allows encoding of bibliographic, performance, and analytical information), with interoperability (can be integrated with digital libraries, archives, and other music-related The Cantus Database is an international network of chant databases that links metadata and musical incipits from liturgical manuscripts worldwide. – COST Action CA 21161, Earlymuse 52 technologies), with preservation and research (useful for digitising manuscripts, encoding historical scores, and supporting computational musicology). The MIRAGE project conducted at the University of Oslo is an example for the use of artificial intelligence in visualising and analysing music — here, the tool developed by Olivier Lartillot visualises a fugue from J.S. Bach’s Kunst der Fuge in real time. 3. Praxis Building on the previous section, the following illustrates how musicologists apply the principles discussed above in their work, and how they situate themselves within the relevant frameworks. 3.1. Organising Beyond the actual digitisation of the media in question, a musicologist engaged with creating data for discovery and use, especially within institutionally-preserved heritage, will capture relevant metadata within appropriate structures. Observing best practices in preservation prevents loss and corruption of data, maintains the integrity of objects and relations between metadata and various – COST Action CA 21161, Earlymuse 53 files, and ensures long-term availability. Care must be taken to ensure that data is maintained and sustainable across changing systems, with new metadata sources integrated, and new highlevel standards adopted and adapted. In data-providing institutions, data is refined, deduplicated, improved manually and by automatic processes on a regular basis. Different schemas and systems are used among heritage institutions, even in different institutions in one country due to subordination to different ministries (science, culture) or due to the specificity of the collections - different systems for paper-based collections (whether printed or handwritten), or for audio collections. The current state is very often a result of legacy decisions that cannot easily be reversed due to the large quantity of manually created data. Passive and active users of this data should be aware of these challenges. 3.2. Annotating Annotations in digital editions represent the substantive claims we make about a musical entity, as opposed to the physical characteristics of the source. Typical categories of annotation metadata include: • Author ascription. Cases of doubtful authenticity abound; the ongoing discussion about the works of Josquin des Prez is an example of this. • Date: based on documentary (external) or stylistic (internal) or evidence. Example: claims about the chronology of the Bach cantatas. Spotlight: Musical Sources in Slovakia The Slovak national group of IAML, in cooperation with the Institute of Musicology SAS and other institutions, initiated a nationwide survey in 2024 to obtain as much up-to-date information as possible about musical sources and the institutions which own them. Approximately 500 institutions (libraries, archives, museums, galleries, religious institutions, etc.) were contacted. This involved distributing a questionnaire on the ownership of music sources, the current state of digitisation, and interest in cooperation or in the professional processing of music collections. Only about 60 responses were received. However, all of the respondents agree that they would welcome a better methodology for processing music sources. Departments often do not have expert staff in the field of musicology. The music sources are not processed in any professional way and the institutions rarely have the financial means to take better care of the music sources, let alone digitise them. Where music sources have been digitised, they are often not publicly available; they are only materials intended for the internal use of the institution or for backup and protection against possible loss (alluding here to both legal issues and problems with storing the data in a central national system). Slovak institutions are clearly facing a major staffing and financial crisis. A selected team of expert staff associated with the IAML national group aspires to provide financial and project support to selected institutions in the future and to create a database of institutions owning music sources in Slovakia. – COST Action CA 21161, Earlymuse 54 • Genre or type: when we say that something is an instance of some class. Example: ‘This piece is a symphony’ (as opposed to other musical genres, such as operas, for example). • Style or technique: the way a piece of music is made and how it is classified accordingly. • Value: that something is good (or good for us), or beautiful; or a good example of its class. • Meaning: claims on the expressivity or semantic determination of a piece. For example, Chaikovsky’s Symphony No. 6 was called ‘Pathétique’ by its composer because of its passionate nature. Systems of annotation and citation are key to this work. We know what it means to quote or cite a particular page of a book. Music scholars also share agreed reference systems and terminologies by which, for example the phrase ‘the exposition of a given Mozart piano sonata ends in measure 58’ is commonly understood independently of any particular edition of that piece. Musicologists are also developing shared ontologies for analytic claims in musical critical assessments, such as OMAC.12 Only when we need to point to a particular manifestation of a piece do we need to think carefully about protocols like EMA or IIIF that will allow us to ‘point’ (in a digital sense) to a specific image. Such ‘image citations’ (see below) could in turn be exploited for machine processing, that is for an OMR-style reading of the semantic content of the notation, or for palaeographical study of scribal practices or identification. Such machine processing also enables citation systems specific to digital objects. MEI encoded scores, for instance, include XML:IDs—digital reference points that uniquely identify every element in a given document. The CRIM Project (the joint work of European and North American scholars) includes notelevel references to scores, which in turn are surrounded with analytic claims by individual scholars. 12 Emilio M. Sanfilippo and Richard Freedman (2022) ‘Ontology for Analytic Claims in Music. In European Conference on Advances in Databases and Information Systems’ An analytic claim by Andrea Puentes-Blanco about a particular composition by Orlando di Lasso, with selected notes shown in red highlights. https://crimproject.org/relationships/2706/. See digital version of the reference via the API: https://crimproject.org/data/ observations/5273/. – COST Action CA 21161, Earlymuse 55 Another example from the same project is concerned with the machine-based annotation of MEI scores,13 in this case via algorithmic identification of cadential patterns in Renaissance counterpoint. The measure/staff/beat positions of the counterpoint is translated into XML:IDs in a corresponding MEI file and then returned to the user to review. This score excerpt is translated as follows: 13 This is accomplished via CRIM Intervals. Ludwig Daser, Ecce Nunc Benedicite Notebook with CRIM Intervals: A Digital Supplement for ‘Lasso takes a Second Look: An Imitation of an Imitation Mass?’ – COST Action CA 21161, Earlymuse 56 For recorded content (audio and video) we also have systems like Music Encoding and Linked Data (MELD), a framework that allows users to create semantic annotations of music concepts and relationships in scores and recordings. Further knowledge is incorporated from external sources using Linked Data, which is also used to match annotation types and contexts to rendering actions displaying the annotations upon the digital score. The MELD Specification in schematic form. Example of MELD-based annotation system, from Lohengrin Time Machine. – COST Action CA 21161, Earlymuse 57 3.3. Enriching The ability to enrich or extend data stems from adherence to FAIR (Findable, Accessible, Interoperable, and Reusable) principles of research data management. Using widely accepted standards and formats not only facilitates reuse of research findings, but also allows subsequent users to enrich these files with additional data. While a project may have individualised needs and have developed specialised formats and schemas for their data, researchers will often strive to adapt to the data model of an existing standard, thus increasing dissemination of their research contributions and allowing future projects to build on prior work. Based on XML as the language of the web scholars of historical and literary documents have developed and used the Text Encoding Initiative (TEI) for some decades now to create dynamic, layered editions, combining data (the texts, including variants, graphical layout) with all sorts of metadata (paratexts about sources, rights, and editorial responsibility) in documents that can be published in forms readable by humans, but also mined for rich information they contain (via various XML protocols like eTree)14. For the last decade music scholars have paralleled this via the Music Encoding Initiative (MEI). Similarly to TEI, MEI is essentially an encoding framework. MEI data can be engraved into music notation using Verovio, an open-source rendering library that can be used in a wide range of environments, including web browsers. MEI can do what no printed edition can manage, representing different versions of a piece, editorial interventions, suggestions, and even analytic annotations or links to images or other remote 14 eTree stands for ‘embedding tree.’ It is an XML element used to represent hierarchical, tree-like data structures within a TEI-encoded document, particularly for representing linguistic or structural analyses such as syntactic trees. See https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-eTree.html. The Verovio music notation engraving library is a versatile tool for rendering MEI encoding scores in a wide range of environments. – COST Action CA 21161, Earlymuse 64 Level Why it matters Example Strategy lIndividual archives, museums, or universities often set their own policies on digitisation, licensing, and data sharing. These may be more restrictive than the law itself, out of risk aversion or resource constraints. A municipal music archive may decide not to digitise recordings even if copyright has expired, due to limited staffing or unclear internal policies. Building relationships with institutional staff (archivists, librarians, legal offices). Their rules often shape your day-to-day access more than EU law. EU directives must be transposed into national law, leading to different rules across Member States. Cultural policy is also largely national, with ministries of culture or copyright offices setting frameworks The 2019 Copyright Directive was implemented differently across Europe. What counts as a text-anddata mining exemption in France may not be identical in Germany or the Netherlands.17 Always checking the national transposition of EU directives in the country where you are working. What is legal in one country may not be in another. Even though the EU has no direct competence in culture, its laws on copyright, data protection, and digital markets strongly affect musicological work. EU funding programmes promote open science and FAIR data. The 2019 Copyright Directive introduced exemptions for text and data mining, but only within certain boundaries, leaving institutions uncertain about how far they can go. Being aware that EU rules create broad goals, but interpretation is uneven. When working in EU-funded projects, aligning your practices with open science policies. Music and research often cross borders beyond the EU. International treaties and organisations (WIPO, UNESCO, Berne Convention, TRIPS) establish global norms for copyright and heritage protection. A recording held in a US archive may be accessible under US ‘fair use,’ but reusing it in an EU-based project may require additional licensing. Conversely, EU projects can contribute to UNESCO frameworks on intangible cultural heritage. Always considering international agreements when collaborating globally. Being mindful of differences between ‘fair use’ (US) and ‘fair dealing’ (EU/UK), and explore UNESCO or WIPO channels for broader advocacy. 17 Text-and-data mining (TDM) exemption: a clause in copyright law that permits researchers to automatically analyse large volumes of digital material (such as texts, images, or recordings) for scientific purposes without needing individual permissions. – COST Action CA 21161, Earlymuse 65 4.2. Current challenges for musicology The dynamic policy and legislative landscape presents a set of challenges that deserve collective attention within the musicological community. As we outlined earlier (see 4.1), this landscape has multiple dimensions: some rules are set at the local or institutional level, others at the national level, while European and even international frameworks also play a role. For researchers, it can be difficult to disentangle these layers. In this section, we highlight key issues that limit the integration of music into broader European infrastructures for heritage and data science. We focus on the European level not because it is the only one that matters, but because EU legislation and policy increasingly set the framework within which national laws and institutional practices operate.18 Reading this section with that in mind can help clarify why certain obstacles arise and where efforts to address them might be most effective.19 Four challenges stand out in particular: 1. Music is rarely recognised as a form of cultural heritage in its own right While music is frequently celebrated as a ‘universal language’ and a symbol of Europe’s cultural diversity, it is rarely articulated in legal and policy frameworks as cultural heritage in its own right. Many of the laws and policies shaping our work were not designed with musicological research or heritage in mind. Existing policies tend to treat music primarily as part of the ‘creative industries,’ focusing on its economic and commercial potential. Initiatives such as the European Commission’s Music Moves Europe programme, for instance, emphasise competitiveness and industry development 18 Here, the findings of the Horizon Europe project FairMusE (‘Fair Music Business Models for a Sustainable European Music Ecosystem’) are especially relevant. FairMusE systematically mapped the regulatory, economic, and social conditions of the music sector, with a strong focus on fairness, competition, and platform regulation. Its outputs provide both an evidence base and concrete policy recommendations that directly touch on the challenges we outline in this chapter. 19 Antonio Vlassis et al. (2025), Deliverable 2.1. Origins, goals and effects of EU law and policy in the online music sector, p. 18. Keywords in relevant EU legislative and policy documents adopted between 2015 and 2023. – COST Action CA 21161, Earlymuse 66 rather than preservation and access to music as heritage. Initiatives like FairMusE confirm this observation. Their analysis shows that EU music policy is overwhelmingly oriented toward industry growth, competition, and digital platform regulation, leaving heritage-related concerns marginal.20 One of the key recommendations to follow from this would be to broaden the policy lens so that music is recognised not only as ‘content’ in a market but also as cultural heritage with long-term public value. 2. Conflicting stakeholder interests Music sits at the crossroads of cultural heritage and commercial industries. As a result, different stakeholders often have competing priorities. Cultural heritage institutions seek flexibility to digitise and provide access, while the cultural and creative industries (CCIs) rely on strong copyright protections to protect their revenue streams. Policymakers, meanwhile, often prioritise the economic weight of industry players, especially in relation to global digital platforms, over the needs of heritage institutions and researchers. These conflicting interests create a policy environment where the reuse of musical materials for scholarship, preservation, or public education is frequently restricted or deprioritised. A major factor exacerbating this issue is that many cultural heritage institutions themselves struggle to navigate the legal landscape surrounding copyright and intellectual property. The uncertainty about what is legally permissible creates hesitation, leading institutions to either delay projects, avoid certain types of audio-enhanced exhibitions, or overestimate legal risks, ultimately resulting in missed opportunities. Museums that are specifically dedicated to music, including collections of musical instruments or particular composers or musical groups, are particularly affected by these uncertainties. Institutions can also impose their own rules for political, economic, or reputational reasons, which may be more restrictive than the law itself. In practice, this means that institutional or platform policies sometimes set the conditions for access, shaping what is actually possible in a project. For example, archives may hesitate to digitise out of caution, or platforms like YouTube may limit certain types of reuse even when research exemptions exist. While this can feel like an additional barrier, it is 20 Ibid., 24–28. – COST Action CA 21161, Earlymuse 67 also an important reminder that law and policy are not the only factors at play: institutional decisions and platform rules are part of the ecosystem we have to learn to navigate. The positive side is that these rules are not fixed: through dialogue, partnerships, and collective advocacy, researchers and institutions can often find workable solutions. 3. Weak legal frameworks for research and reuse Even when music is acknowledged as heritage, the legal tools available to researchers remain weak. Unlike written or visual heritage, musical recordings and related materials benefit from fewer exceptions and exemptions for research and education. The 2019 Copyright Directive introduced textand-data mining provisions, but their uneven implementation across Member States leaves researchers uncertain about what is possible in practice. The impact of these legal uncertainties extends directly to musicologists, whose work relies on access to historical recordings, music metadata, and archival or museal materials. Without legal mechanisms that allow for the responsible reuse of musical heritage, institutions struggle to share materials with scholars, limiting musicologists’ ability to conduct research, create new editions of works, facilitate performances, and contribute to digital archives. The absence of robust ‘fair use’ provisions, combined with complex layers of rights in performances, editions, and recordings, makes it difficult for musicologists to reuse materials. As a result, research outputs (whether datasets, new editions, or digital exhibitions) often remain locked away rather than becoming part of shared infrastructures. The consequences of this legal paralysis are significant. Not only does it restrict public access to Europe’s musical heritage, but it also discourages both institutions and researchers from pursuing innovative projects, especially those involving AI or large-scale data analysis. FairMusE The FairMusE project is an Horizon Europe project dedicated to promoting fairness, transparency, and sustainability across the European music ecosystem, particularly in the wake of the dominance of online platforms and algorithm-driven streaming models. The team shares their insights via reports, as well as podcasts and other outputs. – COST Action CA 21161, Earlymuse 68 What legal rules decide which materials can be digitised and made available as cultural heritage? • Copyright law is the main gatekeeper. In the EU, recordings, compositions, and editions can remain protected for up to 70 years after the death of all of the authors, performers and other creative and intellectual contributors. This limits what archives can digitise and put online without licenses. • Separate rules regulate the protection of privacy in unpublished materials: private content (e.g. private correspondence), personal data, images of private individuals, as well as works that were never published or performed. • The EU cannot legislate culture directly, but directives (e.g. 2019 Copyright Directive, DSM) and national implementations shape what can be digitised for research, preservation, or education. National variations create a patchwork. • Unlike written or visual heritage, musical heritage has fewer clear exemptions, making its integration into infrastructures like Europeana or E-RIHS more difficult. • Even when exceptions exist (e.g. for preservation), archives and museums often hesitate, fearing infringement. This leads to under-digitisation and limited availability. Takeaway → access depends not only on copyright duration but also on how confidently institutions interpret exceptions and whether EU/national frameworks provide clarity. 4.3. Fragmentation and lack of standards Even when materials are legally accessible, fragmentation remains a major barrier. Institutions across Europe apply different digitisation standards, use different metadata models, and interpret copyright exemptions in different ways. This creates a patchwork system where data cannot easily be shared or combined. For infrastructures such as the European Collaborative Cloud for Cultural Heritage (ECCCH) or E-RIHS, which rely on interoperability, this fragmentation undermines the integration of music into broader heritage science initiatives. Researchers trying to work across borders often encounter ‘silos’ of information and face additional costs in harmonising datasets that should, in principle, be reusable.21 21 On this topic, it will be worthwhile to monitor developments around the EU Data Act (active since September 12, 2025) and the forthcoming Data Union Strategy. – COST Action CA 21161, Earlymuse 69 These challenges cannot be solved by individual researchers alone. They require collective responses: stronger representation of musicology in European heritage debates, clearer legal frameworks that balance public and commercial interests, and the development of standards and infrastructures that account for the particularities of music. FairMusE shows that these collective responses are possible: its future agenda explicitly calls for a ‘fairer music ecosystem’ that balances industry competitiveness with cultural sustainability, and opens up opportunities for researchers and heritage professionals to be part of this process. In the final section of this chapter, we will look at future developments and opportunities for us to engage and address these issues, both as individual scholars and as a musicological community. 4.4. (How to navigate) challenges you might encounter Researchers often encounter legal and policy questions in three main areas: access, use, and circulation. Below we outline what this means in practice, with short examples you might recognise from your own projects. The digitised holdings of the historical sound archive of the German National Library are searchable in the online catalogue (though only accessible via listening booths in their physical reading room). – COST Action CA 21161, Earlymuse 70 4.4.1. Access: what can be digitised and made available? Access depends mostly on copyright and institutional rules. Can we obtain the materials we need? Challenges we face are: • Copyright and related rights: musical recordings and scores may remain protected for decades, creating uncertainty about what can be accessed freely. • Fragmentation: archives and institutions vary in their access conditions, licensing terms, and digitisation priorities. This creates a ‘patchwork’ of resources rather than a coherent European landscape. • Digital divide: while some collections are well digitised and openly available, others remain locked behind institutional walls or exist only in analogue form. Example 1: Public domain scores. If you want to digitise an 16th-century manuscript from your library, you can usually do this freely because copyright has expired. Example 2: Recordings of the same music. A 1970s LP of that composition may still be protected. Even if the composition is public domain, the performance and the recording have their own rights. Digitising and sharing that without permission can be a problem. Example 3: Archive restrictions. Some archives limit digitisation because they do not know the copyright status or do not have the resources to manage permissions. In practice, this means you might get access on-site, but not be allowed to put files online. Takeaway → ask about rights status and if in doubt, check with librarians or rights offices before planning a digital project. Access depends not only on copyright duration but also on how confidently institutions interpret exceptions and whether EU/national frameworks provide clarity. 4.4.2. Use: what can be analysed with digital tools? Key challenges around copyright, data protection, and platform rules include: • Licensing restrictions: some digitised collections allow only personal study, not text-and-data mining or computational analysis. • Unclear exceptions: the EU’s DSM Directive (2019) introduced text and data mining exceptions but their implementation varies across Member States, creating uncertainty for cross-border projects. – COST Action CA 21161, Earlymuse 71 • Standardisation gaps: lack of consistent metadata and formats hampers large-scale analysis and reuse across datasets. Example 1: Text and data mining. The 2019 EU Copyright Directive introduced exceptions that allow researchers to mine text and data (TDM). This means you can, for instance, run algorithms on large corpora of digitised scores. But recordings are more complex: the Directive left many grey areas, and not all Member States interpret it the same way. Example 2: Personal data in archives. Suppose you are working with oral histories of performers. GDPR rules mean that names, addresses, or even rehearsal comments may count as personal data. You may need consent before analysing or sharing them. Example 3: Working with streaming platforms. If you want to scrape data from Spotify or YouTube to analyse audience behaviour, platform terms of service (not just copyright law) decide what you can or cannot do. These rules are often stricter than EU law and can shut down projects if ignored. Takeaway → having the files does not automatically mean you can use them for digital analysis. Check copyright, data protection, and platform conditions before starting. 4.4.3. Circulation: what can be shared across borders? Can we share research outputs like datasets, analyses, or digital editions with the wider community? EU directives and regulations try to harmonise rules, but national copyright laws differ, making crossborder projects tricky. International treaties like the Berne Convention also apply. For example, a recording you can legally use in the US under ‘fair use’ may still require licensing in Europe. Challenges include: • Rights over derivative works: even when scholars create new datasets from existing recordings, rights holders may claim restrictions on redistribution. • Institutional silos: research outputs are often stored in project-specific repositories, making long-term accessibility uncertain. • Policy expectations: EU programmes increasingly require open science and FAIR data principles, but meeting these demands is difficult without clear legal pathways for musical materials. – COST Action CA 21161, Earlymuse 72 Example 1: Different rules for ‘fair use.’ In the US, using short audio clips in teaching or research may be covered by ‘fair use.’ In Europe this often still requires a license, even for research. A dataset legal in one country may not be legally usable in another. Example 2: EU vs. national differences. The EU tries to harmonise rules, but directives are implemented differently in each Member State. A project involving institutions in Germany and Italy might discover that what is permitted in one country is blocked in the other. Example 3: Sharing metadata. You may think metadata is ‘safe’, but even sharing basic details like recording dates or performer names can raise legal questions under copyright or GDPR. Some institutions restrict metadata exports, even though they seem like harmless facts… Takeaway → circulation is not only about your research outputs but also about the underlying data. Plan early if your project involves partners across countries, and check national as well as EU rules. 4.4.4. Navigating the challenges: checklist for research projects The legal and policy environment can feel daunting, especially if you are just starting out in digital or heritage-focused musicology. Still, it is worth our effort. Most projects can move forward once you know where to look for answers, who to ask for help, and how to plan ahead. Navigating these hurdles not only protects your work but also strengthens your ability to advocate for better access and, hopefully, more research-friendly legislation and policy in the future. A good first step is to map out which levels of law and policy apply to your project. Identify what kinds of materials you will be using (manuscripts, recordings, editions, metadata, oral histories) and check which legal and policy frameworks apply. Being aware of this avoids dead ends and helps you bring questions to the right people. In general, we suggest seeking support (librarians, archivists, and research data specialists often know the latest about copyright and licensing) and staying pragmatic (many issues have workarounds; you do not need to abandon your project at the first obstacle). A local institution with a legal department or a colleague who specialises in law might be able to help you map the national context, both for yourself and for joining cross-border projects. Establishing contact with, or acquiring knowledge of, the national authors’ rights societies (which play a major role in music-related copyright issues) can also be very valuable. You can find information about the different societies and what they do here (scroll to the bottom of the page for a directory by country). – COST Action CA 21161, Earlymuse 73 Key aspects to keep in mind: • Copyright status: always check whether the material you want to use is still protected. Manuscripts and scores may be in the public domain, but recordings and edited editions often remain protected for much longer. Librarians or research data specialists can guide you through this. • Licensing and permissions: critical editions, recordings, or even performance notes may require licenses before you can reuse or publish them. Remember: using a work in private research does not automatically give you permission to include it in a publication or dataset. • Digitisation: if you plan to digitise sources, make sure this is legally allowed (even if you are doing so only for preservation). Some institutions are risk-averse and may limit what you can digitise without formal clearance. • Metadata and analysis: using metadata for computational analysis can be tricky. Some institutions treat metadata as freely reusable, while others impose restrictions. Confirm what is allowed before investing in a large-scale project. • Performances and recordings: if you record performances as part of your research, you will usually need consent from performers, composers, and rights holders. Sharing these recordings later often requires extra permissions. • Publications and open access: many funders now require open-access publication of research and data. Check your institution’s policies and make sure that any musical examples (score excerpts, audio clips, etc.) comply with copyright. • Cross-border collaborations: if your project involves partners in different countries, be prepared for different national implementations of EU law. What is possible in one country may not be legal in another. Review agreements carefully, especially for data sharing. To help you along, we provide you with a practical list of questions and actions to consider. Think of it as a companion you can return to whenever you are planning, conducting, or publishing your research. This checklist will not answer every question, but it will help you spot red flags early, involve the right people, and make your research more sustainable. And over time, our collective awareness of these issues strengthens musicology’s voice in European heritage and policy debates. Checklist for musicological research projects Before you start: • Map out which levels of law/policy apply (institutional, national, EU, international). – COST Action CA 21161, Earlymuse 80 home in a workplace that looks increasingly like ‘big data’ science and social science, with extended networks of collaborators co-authored publications, and standardised (or at least consensus-based) methodologies that lead to (near) reproducible results. Preparing ourselves (and in turn, the next generation of musicologists) for this kind of work will require some re-education and reformulation of curricula. We need to foster literacy in new ways of gathering, enriching, combining and interpreting music sources, which are increasingly understood as data. As technologies evolve at an ever-increasing pace, sharing knowledge and educational resources both vertically (to students) and horizontally (between researchers) is the only sustainable option. Important is also to reach out and connect with related fields like Music Information Retrieval and Music Computing29, where high technical expertise exists, and who are prime target audiences for (historical) musicological data of high quality. Such work will be particularly important for the success of institutions and scholars with only modest technical infrastructure. And so we should be intentional about the creation of: • curricula: modular assignments that can be adapted for different data sets and institutional contexts, including sharable code that can be used to work with various data sets in project-based assignments; • workshops with experts to train instructors, their assistants, and students, both within and across institutions; • networks to exchange expertise and ideas (like a ‘digital musicology skills network’); • internships or even ‘mid-docs’ for emerging scholars, perhaps modelled on programs like the ones offered via the Price Lab for Digital Humanities at the University of Pennsylvania. To date, musicological research is largely disconnected from the cultural and creative industries (CCIs). While there are numerous points of interaction, long-established ways of engaging with the CCIs, such as CD booklets, are disappearing. It is rarely the case that people from the creative or cultural sector are in dire need of musicological expertise; rather, it is usually up to musicologists to try to establish links or make their work known more broadly. 29 It is a research field that combines scientific, technological, and artistic methodologies in order to understand, model, and generate sound and music through computational approaches. – COST Action CA 21161, Earlymuse 81 This is not true for music per se. The music industry is massive (one of the largest globally), having pivoted quickly in the last 20 years from the material mediation of the CD to the world of streaming. In this new landscape, music means almost nothing on its own, but is instead refracted through the behaviours of hundreds of millions of users, at least since the advent of Matrix Factorization. Such work has required a host of computer and data scientists to manage and organise. However, these specialists, as Nick Seaver30 and Liz Pelly31 have recently argued, rarely have a background in musicology, and there is an evident skills gap that needs to be bridged. Apart from educating ourselves and the next generation of musicologists to become visible and appealing to the music-related industries, we see several further points of connection, many of which are already underway: • Using music data in public music education • Using music data in multimedia performances • Using music data for heritage sites But what do we musicologists actually bring to the table? Arguably, it is the quality of (meta)data that musicology can provide. However, we should remember that in the big data age, quality loses to quantity - at least in principle and from the perspective of the music industry. Thus, it might be helpful to shift the gaze from the very big players to somewhat smaller ones, with whom fruitful interactions seem more likely. National museums and galleries increasingly explore new formats for their exhibitions where music (or sound more generally) plays an important role for user immersion. New European Bauhaus policy and funding initiative, aiming at sustainability, social inclusion and aesthetics, seems tangential to this particular trend. 30 Nick Seaver (2022), Computing Taste. Algorithms and the Makers of Music Recommendation. 31 Liz Pelly (2025), Mood Machine. The Rise of Spotify and the Costs of the Perfect Playlist. – COST Action CA 21161, Earlymuse 82 5.2. Linking and accessibility The FAIR principles are a well-established set of guidelines in the world of Open Science (see for instance the OSCARS project). But for music (and musicology) these principles in many respects remain an elusive but vital goal. Part of the challenge is simply the vast (and often strange) data that are part of the musicological domain. European symbolic-graphical notation forms a kind of ‘interoperable’ standard by which musicians can communicate with each other. But flat, graphical files (like PDFs or images) are hardly amenable to the new modes of reading and interpretation we will need to foster in the decades ahead. Meanwhile, we lack any consensus on a specifically logical/digital means by which we can encode a musical work. Instead we have things like MIDI files (which originated as a system for controlling electronic instruments), Music XML (a kind of de facto interchange standard between music notation software), or (for about the last decade) Music Encoding Initiative files (like the TEI standard, with a rich capacity for scholarly annotation and metadata), among others. Transforming one into the other is no trivial task. Nevertheless, LLMs already have a surprisingly good representation of these structures, and can be of considerable assistance in converting one form to another.32 In brief: interoperability that minimises information loss remains a key priority for work with Music as Data. Another important challenge for FAIR datasets is found on the other side of the equation: data about Music. Of the encoding standards noted above, only MEI contains robust capacity for linking scores to the standard identifiers (like VIAF) for the composers and editors who create them, and thus for managing the layers of intellectual property that are a key dimension of all ‘responsible’ (in the FAIR sense) of work with the performing arts. Yet even if we could reliably connect the texts with their makers, we would still face an even more difficult problem: the thorny Work Concept itself. RISM, as a resource founded to make a systematic inventory of sources, has so far managed to side-step the challenge of creating persistent worklevel identifiers for musical compositions themselves, which remain in many respects convenient abstractions (tests with the exclusive use of printed thematic catalogues of composers are underway). Humans are prepared to understand the many renditions of (for example) a Bach partita: Bach’s autograph manuscript copy, the edition in the Neue Bach-Ausgabe, the copy on your piano, the CD 32 See Richard Freedman and Daniel Russo-Batterham (2024), ‘Musicology in a Time of Technological Transformation’. – COST Action CA 21161, Earlymuse 83 on your shelf, and the many MP3s on Spotify as expressions of one and the same artistic idea. But for machines, ‘finding’ these manifestations automatically across score archives and databases remains a serious challenge. Do we need formal ontologies and sets of identifiers for musical works as a way of connecting them with scores and the vast array of metadata (biographical, geographical, codicological, institutional) that help us put them in context? Professional librarians and database designers would probably say yes. Yet, the success of public datasets like IMSLP, or ChoralWiki, or even various Kaggle compilations33 reminds us that while such loosely curated resources generally do not fully conform to our scholarly standards, we should not simply dismiss them as irrelevant. On the contrary, they highlight what many users value: easy access over complicated metadata models, broad coverage instead of deep contextual information, free access over pricey but beautiful printed texts. The solution to this challenge of linking ‘music as data’ with ‘data about music’ will thus likely involve work from two directions: the development of formal schemes for understanding and exposing metadata in a structured way (through published ontologies, unique identifiers, and carefully structured APIs or other endpoints); and experimentation with LLMs (see below) as unsupervised ways of mapping musical knowledge (connecting people, places, roles, events, and even ‘works’ through their contextual proximities rather than through any a-priori categories). 5.3. Reading and Writing New technologies will provide us with new layered and complex ways of reading and writing musicological texts. In fact, the notion of a musical text will likely broaden further to include corpora of digital music encodings, databases, linked and interconnected repositories, as well as performance recordings. What might these new modes of interpretation and knowledge creation look like in the decades to come? Reading, broadly understood as distilling knowledge from information represented in some text, will expand to include data-analytical skills necessary to ‘read’ lines of computer code, numerical 33 Kaggle compilations refer to curated collections of datasets available on the Kaggle platform, often used for data analy sis, machine learning projects, and competitions. These compilations bring together diverse and well-structured data, including music-related datasets, which researchers and practitioners use to develop, test, and benchmark computational models and algorithms. – COST Action CA 21161, Earlymuse 84 data and statistical information. Digital data models of musical (meta) data are becoming increasingly complex and may require abstraction to remain intelligible (readable), for which artificial agents may be leveraged. We can no longer assume we have seen the entirety of the data, even with larger teams of researchers. Instead, scholars will increasingly need to apply pattern-finding and summarization techniques to properly read data from a distance. Retrieving patterns, however, is only the first step in further musicological inquiry or analysis. These patterns need to be connected back to the contexts from which the sources or data stem. This knowledge extraction will likely happen in increasingly coordinated ways between humans and machines, e.g. when scholars formulate and refine research questions with the help of a LLM (‘prompting’) in a feedback loop. Growing amounts of data will allow shifting between scales of reading (from close to distant) and enable exploration of new areas yet to be studied . For example, we can attempt to trace the transmission of musical patterns through time and space, or infer communal practices from document distributions. The emergence of these modes of reading and writing will in turn depend on domain-specific tools, ‘Small Language Models’ or Retrieval-augmented generations attuned to the unique features of music—the very mixture of Music as Data and Data about Music mentioned above. Viewed from the work of Data Science, the musicology domain is a ‘low resource domain’: we might think of a corpus of 100,000 encoded scores or theoretical writings about music, or biographical notices about musicians might seem to be Big Data, but from the standpoint of bioinformatics (or Facebook posts), they are decidedly small. This need not be conceived as a disadvantage of our object of study (and, consequently, our field). Low-resource domains like musicology with ‘littleish’ data can prove extremely valuable in evaluating and improving existing information processing systems. The challenge they pose in their relative sparsity is that they bring along a considerable amount of uncertainty, ambiguity, and irregularity. Robust and reliable machine-learning frameworks that are supposed to approximate intelligent agents need to be able to deal with unwieldy data as well. For example, newcomers to machine learning will quickly encounter datasets for classification such as hand-written digits (MNIST) or clothing items from Zalando (fashion MNIST). Both are relatively small (60k+10k images with 28x28 pixels), but are relatively easy to classify because they have clear – COST Action CA 21161, Earlymuse 85 but relatively few underlying categories (10 digits or fashion items, respectively). Imagine picking 70k recordings of pieces of music (different pieces, not performances). How would we expect that these would cluster together if only 10 categories are allowed? The point is not that musicologists need to come up with clearer categories or ontologies. Music taxonomies like genres (and sub-genres, and sub-sub-genres…) are inherently complex, overlapping, and non-hierarchical, which makes automated classification difficult, given that these categories will commonly not contain tens of thousands of examples. Recent decades have seen a growing number of projects that produce digital editions or digital corpora of music, and it is impossible to provide an exhaustive list here. But even if they are using the same or similar encoding formats, combining them or filtering a particular subset from all of them is to date unfeasible or highly time consuming. The solution is not a centralised system (like RISM) that hosts (ideally) everything, but a distributed network of interlinked and interoperable datasets that can be queried as needed. With the right Little Language Model, we thus might be able to imagine several new modes of machineassisted reading of music as data and data about music. For instance: • interact with structured data using natural language (making the information in catalogues and databases accessible open to new forms of interrogation; Anthropic has recently announced the Model Context Protocol that aims to make exactly this possible), • make structured data from narrative texts (return entities and relationships from some historical texts, OCR versions of some historical newspaper, or theoretical writings; these as a way of helping advance the creation of catalogues and databases), • pose qualitative questions of historical writings (ask historical texts about ‘novelty’, ‘beauty’, ‘order’ ‘harmony’, etc.), • query musical scores and recordings (ask a question about a score; align recording and score; discover patterns, large and small, distant and near). – COST Action CA 21161, Earlymuse 86 Scholarly writing will likewise take new forms. Instead of fixed texts published in print, we foresee a growth in interactive, data-driven publications. These will both make our work reproducible and provide access to a larger audience, e.g. through data visualization or sonification, instantaneous automated translation, etc. The data inform write-ups of research; however, those are not just academic monographs and articles, but also wider distribution of research results through museum exhibitions, books for a wider audience, recordings, films, radio shows, productions, YouTube videos, etc. Scholarly citation practices may change in various ways. The range of items ‘acceptable’ to cite might broaden from published books and articles to online resources, datasets, and software tools. Persistent and versioned identifiers are therefore necessary to trace exact versions. The digital domain allows for non-linear and dynamic texts (e.g. Wikipedia). Large amounts of data might lead to empirically back up musicological information online by relating statements directly to the sources (links) or databases/sets (statistics). Authorship will also transform. Traditionally, musicologists tend to produce single-authored texts. We imagine that not only will co-authorship become more prevalent, new and different roles of people involved in producing written research outputs will emerge, similar to practices already established in the (social) sciences (e.g. data collection and cleaning, data analysis, manuscript writing, etc.). We need to ask ourselves how people who have contributed to our work (e.g. students encoding scores during coursework) will be appropriately acknowledged and potentially remunerated. MULTISCORE, a project of the University of Alicante. – COST Action CA 21161, Earlymuse 87 Digital musicology and artificial intelligence now play a fundamental role in all aspects of the field, from source description to work editing, from prosopographical data to musical information which informs both historical narratives and analyses. The integration of digital tools allows for more sophisticated analysis and visualization of musical data, enhancing the depth and accessibility of research. Digital projects often prove to be effective vectors of collaboration, extending beyond institutional boundaries and geographical limits—but they demand adjustments to our horizon of expectations. 6. New models of publication Do current forms of publication in musicology align with current requirements regarding data production, interoperability, integration of multimedia elements, and accessibility? Do these publications reflect the diversity of expression modes in musicological approaches, ranging from a readable text to an audio-visual sequence? In a previous report, we explored avenues for the future of musicological publications.34 Here, we address the issues for imagining and designing a new model of musicological publication that pushes the boundaries of traditional publications.35 As already discussed in section 1 of this report, Europe possesses various infrastructures that support the values and principles to which we adhere: open, participatory, and resolutely inclusive science. For example, the European Open Science Cloud (EOSC) offers an ideal framework for the deployment of new publication models. The Social Sciences & Humanities Open Cloud enables the deployment of tools that will prove necessary for the realization of publication models in which all data contributing to musicological knowledge can be combined. OPERAS offers opportunities for publications of a new kind. A new model of publication necessarily implies digital or hybrid publications that: • are open access (a key aspect in the European landscape of science) • integrate technologies developed over the last years/decades 34 Philippe Vendrix, ed. (2024), ‘Musicological Publications in the Digital Age: Charting a New Course’. 35 See also, for example, Directorate-General for Research and Innovation (2019), ‘Future of Scholarly Publishing and Scholarly Communication - Report of the Expert Group to the European Commission’. – COST Action CA 21161, Earlymuse 88 • incorporate multimedia elements (audio, video, scores) • engage with different users • are more accessible • address inequalities • enrich research as much as practice and pedagogy This new publication model aims to reach diverse audiences: • researchers, musicologists and performers from all backgrounds • professionals dealing with music-related materials in heritage institution • stakeholders in the cultural and creative industries • decision-makers, from research institutions to European organizations • the general public consuming music All of these topics will be addressed in this document. The aim is not to describe a closed model, but rather to outline perspectives for development in the coming years. At the very least, we hope to have provided elements that will allow everyone, in a creative and innovative way, to integrate musicology into the dynamics of an open and participatory science. 6.1. Limitations of traditional publications The results of musicological research have traditionally been published in monographs, edited volumes and specialist journals. Textand notation-based publication is embedded in the institutions in which musicology is located as a professional endeavour, both in terms of education and career progression: (peer-reviewed) articles and monographs are currently essential outputs for most scholars engaged in musicology. However, as musicology develops to embrace new technologies and methodologies and to investigate sounds and practices that existed in semior non-notated forms, it is creating more inclusive and heterogenous outputs. Traditional publication modes can expand and transform to meet the challenge of delivering these outputs in forms that can be accessible to the widest possible audience. – COST Action CA 21161, Earlymuse 89 6.1.1. Physical, linguistic, and economic limitations In most areas, printed publications are still the standard, although simultaneous or exclusive distribution as e-resources is becoming more and more common.36 For example, a search for the topic ‘musicology’ in WorldCat returns 34,531 hits. Filtering by media type shows that there are 7,200 printed books compared to 3,800 e-books, and 1,000 printed journals compared to 820 electronic journals (publications that appear in both forms are counted twice). The figures for open-access publications are much smaller: only 488 books and 106 journals are listed here.37 The number of openaccess journals listed in the Directory of Open Access Journals, which includes only journals that meet certain standards, are even smaller. Here, 105 journals on the topic ‘Music and Books on Music’ are indexed, of which only thirty-six are from member states of the European Union, with ten from the UK. However, a closer look reveals that, for example, none of the ten UK journals come from the field of music history. The same is true for the three German journals indexed and for at least three of the eight Spanish journals. Not so long ago, the leading journals were still primarily printed journals. This has changed in recent years. Fee-based online access, occasionally with a moving wall (e.g. via services such as JSTOR), has established itself as an additional form of distribution, as has the phenomenon of transforming individual articles into open access (usually through the payment of article processing charges [APCs]; see further below). However, these services often provide access only to the ‘flat’ PDF of a printed article, rather than any enhanced version created on the original journal’s website. In Germany, for example, this applies to Die Musikforschung (the journal of the German Musicological Society)38 as well as for other comparatively high-circulation journals such as Musiktheorie: Zeitschrift für Musikwissenschaft, Die Tonkunst: Magazin für klassische Musik und Musikwissenschaft, Archiv für Musikwissenschaft and Musik und Ästhetik. 36 ‘Both print and html’ is already explicitly a part of all modern content management systems for publishers. See University of Chicago (2024), The Chicago Manual of Style, 18th edition, section 1.87. 37 This is not a European phenomenon only; the numbers given are not limited to Europe. The situation in the US seems to be even more problematic, as reported in Rachel E. Scott and Anne Shelley (2022), ‘Music Scholars and Open Access Publishing‘. 38 It should be mentioned however, that the Musikforschung is provided in Open Access via musiconn.publish with a moving wall of thirty-six months. – COST Action CA 21161, Earlymuse 96 • Danielle Fosler-Lussier, Music in America’s Cold War Diplomacy, digital companion (website and database to the book). • RicercarDataLab, Renaissance Data Projects, collection data from projects conducted over the past twenty years on Renaissance individuals and works. • European Ars Nova, Multilingual Poetry and Polyphonic Song in the Late Middle Ages, digital platform (database, publications and audio recordings). • CRIM project, Cyclic Renaissance Imitation Masses, digital research environment (advanced digital tools to investigate the Renaissance imitation masses;focus on annotation tools and analysis). • Renaissance Society of America, RSA Digital Resources, online repository (digital work regarding the Early Modern period.) . • Echoes from the Past project, Unveiling a Lost Soundscape with Digital Analysis, development and application of OMR and MEI to early chant repertory from medieval Portugal. A new format of publication could also foster collaborative and interactive research.41 Traditional, narrative-based publications in musicology are most effectively written by a single person or a small group of people only, and thus do not allow for the integration of multiple points of view. Many interdisciplinary projects, projects that include audio and/or audiovisual contributions and those that rely on citizen science, to name but a few examples, cannot be presented adequately by means of ‘traditional’ musicological publications, despite being important and promising elements of current musicological work. Furthermore, a publication platform could not only accommodate such projects, but also encourage the exploration of new ways in which we could imagine and organise musicological scholarship. 41 See, for example, Janneke Adema, Simon Bowie, Marcell Mars, Tobias Steiner (2022), ‘Books Contain Multitudes: Exploring Experimental Publishing’. Echoes from the Past (Universidade Nova de Lisboa) offers tools for analysis and guides for MEI encoding. – COST Action CA 21161, Earlymuse 97 6.2. Leveraging technological advancements for a new model of musicological publication New technologies of writing have always complicated the relationship between authors and readers. This is especially true of the performing art of music. From the beginnings of Western musical notation to the advent of the printing press, sound recording, and now the digital domain, new technologies of transcription brought about means for controlling the effects and purposes of music, even inaugurating a new sense of it as intellectual property. Each was a ‘new medium’ of its day, and each brought with it new ways for composers, performers, and listeners to interact around musical ideas. What might such change mean for musicology, and the musicologists who make it? We once viewed print as the durable medium through which we put our best ideas before colleagues and the wider musical public. Now digital texts erode critical authority while simultaneously heightening the need for us to take responsibility for the claims we make. These days almost anyone can create a digital archive to store and distribute the ‘flat’ graphical symbols of standard musical notation (a PDF file will do). Indeed, resources like IMSLP or the Choral Public Domain Library are already heavy with hundreds of thousands of both historical documents and home-grown editions. The authoritative Urtext and error-laden mess are offered up as interchangeable versions of the truth. Making musicology in a digital age will require us to think carefully about musical texts, in all their richness, from original source to computer screen, and from traditional prose text to recorded performance. To this we add new ways of understanding music as data (through layered, encoded scores) and data about music (the richly layered metadata that puts music in the context of how it was made and heard). What kinds of publications will be supple enough to support this kind of work and the new modes of interpretation and argumentation that will emerge around it? We might see all of this through two pairs of key concepts—ideas that both complement each other and that will serve as focal points for effective policymaking. The first of these is ‘delivery and discovery,’ which are deeply related to each other. The second pair is ‘sustainability and innovation,’ which are also joined, but in a way that is sometimes contradictory. For each, we detail some important matters of policy and practice. – COST Action CA 21161, Earlymuse 98 6.2.1. Delivery and discovery Musicological work will increasingly depend on a very wide range of ‘evidence’. Surely, we will still have a need for traditional kinds of texts familiar in books and print journals: textual documents, graphical scores, and images. But alongside these we will need now to find ways to deliver sound and video recordings (demonstrating different interpretive approaches to the same text), richly encoded logical representations of scores (an XML document with different layers of editorial intervention, for example), or complex datasets and the tools to interpret them (e.g., a network of musicians active in a particular time and place, and the code used to make the connections legible). Musicology is not alone in making use of these kinds of digital texts. But there is a pressing need for the creation of some new platforms that will allow scholars to engage the full range of evidence and argumentation suited for a digital age. These include: • Digital publication and content management systems that will support rich multimedia content, including colour and high-resolution images, audio and video recordings, along with rendering systems for XML encoded scores and interactive digital tools (such as networks or visualizations). It is also the case throughout the natural and social sciences that authors include data sets and even the code for algorithms used in the interpretation of their data as supplementary materials, which are nevertheless not explicitly part of the Version of Record (typically a DOI and associated PDF and HTML). Associated action item: sponsor work to advance the development of such platforms, perhaps as part of a collective initiative that would share tools and hosting platforms. • Workflows and Best Practices for editors, authors, and publishers, so that born-digital work can make its way reliably from a scholar’s desktop to a reader’s screen in efficient and reliable ways. The Chicago Manual of Style details the modern publication system, from author manuscript, through XML encoding of texts and associated figures and examples, through publication in PDF, HTML, e-book and other forms. These methods should be common knowledge, understood across three key constituencies: –Individual scholars need to understand this process so they can anticipate how they can think of the full range of materials available for their arguments. Understanding of the digital approach to content management and publication should be a part of graduate training, and the object of discussion during colloquia. Associated action item: make experts available to visit classes and conferences. –Editors need to have reliable ways of managing content and to avoid costly delays or confusion in the production of complex texts. This means ongoing training for the professional team involved in production, and various kinds of ‘on-boarding’ training for the scholars charged with the solicitation and review of materials for publication. Digital and multimedia content (are not simply ‘illustrations’; they are part of – COST Action CA 21161, Earlymuse 99 the argument, and editors should understand how to seek appropriate peer review for born-digital work. Associated action items: create handbooks of best practices, and workshops for editors and professional staff. –Publishers, whether traditional commercial houses or the various national and disciplinary bodies should have access to professional advice on how to produce such work. Associated action items: for those who already have robust publication management workflows, there should be technical support that would help them implement some new tools, and document them for others. For small journals as yet unfamiliar with modern content management systems, a ‘turnkey’ system might be developed that would let them manage their work in a modern way (and move gradually towards some of the new tools and possibilities). Consider ‘partnerships’. No less important than the texts themselves (in all their variety): how will users find the objects, images, scores, and data they need? Digital Object Identifiers (DOIs) are of course now common practice in the world of publishing. But to this we need author-level identifiers (such as ORCIDs). But what other kinds of metadata (and endpoints for accessing them) are needed to ensure discoverability by both humans and machines? The development of literacy and capacity with the following concepts seems crucial: Application Programming Interfaces (APIs) for databases makes the structured data that sits behind the forward-facing portal for humans discoverable by machines. In their simplest form these APIs could be JSON or other standardised ways of publishing information. Recently, HumaNum has created a strong solution for accessing information across databases by leveraging high-level terms in differently structured APIs. The operation of such ‘federated’ searches will depend on having well-documented APIs for every project. See, for example, the clear explanation of the system used by the Digital Image Archive of Medieval Music Manuscripts (DIAMM). For example, compare these representations of the same information: – COST Action CA 21161, Earlymuse 100 We note that accessing the kinds of SPARQL endpoints used in many projects are very challenging. For humans: https://www.diamm.ac.uk/sources/117/. For machines: https://www.diamm.ac.uk/sources/117/?format=json. – COST Action CA 21161, Earlymuse 101 Linked Open Data has also been a dream of the Semantic web for many years now. There are, of course, some solid systems in place that help us understand biographical, geographical and institutional entities and their relationships (as we think about ‘who studied with whom in which place and at what time’). See, for example, Cultures of Knowledge, a major LOD project. However, the state of the art for intangible heritage (like musical works) is still relatively weak. Much more could be done to develop reliable ways of pointing to such abstractions in ways that would connect scholarship in a digital age. Artificial Intelligence will probably work both to advance and disrupt how we ‘discover data about music’. In particular, the creation of ‘Small’ Language Models (as opposed to general models of artificial intelligence) could help us interact with databases using natural language, or help researchers develop structured data from narrative and encoded versions of musical scores.42 Neural networks are already widely used in numerous applications. A good example is the End-to-end Audioto-Score (A2S) tool, created during the project Multiscore (Alicante University), which aims to derive a score that represents the music content of an audio recording in a single step through the use of Convolutional Recurrent Neural Networks. 6.2.2. Sustainability and Experimentation What is sustainable in the digital age? This is not a trivial question with respect to musical scholarship. Many journals and academic publishers rightly think of the PDF as the ‘version of record’ for their work—a kind of flat representation that emulates the layout of a physical book. However, recent highly 42 Richard Freedman and Daniel Russo-Batterham (2024), ‘Musicology in a Time of Technological Transformation‘. A Transformer Approach For Polyphonic Audio-To-Score Transcription: learning scheme. – COST Action CA 21161, Earlymuse 102 publicised data disasters (like the cyberattack on the British Library’s online information systems; or the cyberand legal threats to the Internet Archive,43 which hosts a plethora of material) have caused many to wonder just how safe and sustainable digital repositories actually are. Clearly, we need to assure ourselves—as scholars, as publishers, and as the librarians who maintain the archive of musicological scholarship—that digital materials will be available in a permanent way. Sustainability is of course also related to access, and the pressing need for open ways of making scholarship available in perpetuity. Musicological work must also be open, since it cannot be a public good if it is not freely available to all. But to this we add a new complexity: how to balance innovation with sustainability? If a journal article includes audio or video content, what does the version of record look like? And how much more does this complicate the possibility for new kinds of interactive, ‘born’ digital materials? In brief, how do we ensure that we create things that are the ‘first of their kind’ and not ‘one of a kind?’ A periodical published by Springer dedicated to multimedia publications provides some insights: Multimedia Tools and Applications. To date, there is no fully satisfactory solution. 6.3. Addressing inequalities through new publication models European music scholarship, while rich and diverse, is traditionally concentrated in specific geographic regions and institutional contexts, hence, the dissemination of research in the field has long been influenced by significant inequalities. This section explores how a new model of musicological publication platform can address these inequities. 6.3.1. The landscape of musicological publishing: existing inequalities Especially scholars based in less well-funded institutions or working in less widely-spoken languages have faced barriers including: Financial inequality: Scholars from less well-represented countries and institutions often lack the financial resources required to publish their research in established academic journals or through prestigious publishing houses. Funds are necessary for covering publication costs, editing services, or 43 The copyright lawsuit concerning the Internet Archive, which amongst others hosts the WayBack Machine that archives internet pages and hosts, has only recently concluded. The effect on the field of musicology, certain branches of which rely strongly on the site’s archives, is unclear at the time of this writing. – COST Action CA 21161, Earlymuse 103 open access fees, as the current open access publishing model perpetuates inequalities. The financial resources available for musicological projects can be overall much lower than in larger countries, which also affects the quality of research being published. Linguistic inequality: The vast majority of musicological research is published in English, which, while an international academic standard, presents a barrier to non-native speakers. Musicological research is often grounded in local or regional contexts that involve non-English primary sources, and a significant portion of the scholarship is published in languages such as Italian, French, German, or Spanish. However, these works may remain largely inaccessible to scholars in other parts of the world, limiting the scope of musicological research and impeding international collaboration. Funds for English language editing or translations are generally lacking. Recent developments in LLMs and AI are promising (though not to the same extent for all languages), but open new questions of intellectual property and research ethics. Representational inequality: The traditional publishing landscape remains dominated by a handful of wellresourced publishers and journals, many of which focus on the study of Western European music traditions and publish in English. Scholars from regions with less institutional support and smaller languages find it difficult to access prestigious venues, which limits opportunities for researchers outside of Western academic hubs. Hence, their topics are less well represented internationally, and unequal representation can result in unequal career development. Social and ability inequality: The basic principles of universal design - creating products that allow maximum participation regardless of ability - contend that good design benefits all that use the product. Universal or barrierfree design should encompass equity, flexibility, perceptibility, tolerance for error, and simplicity. It should also anticipate feedback from users and allow for future adaptability in response to that feedback. This last principle of responsiveness is crucial to longevity as well as accessibility. Universal and barrier-free design the social, cultural, intellectual, and physical accessibility to the understanding, practice, and enjoyment of the published products of historical musicology by the widest possible constituency of scholars, readers, and audiences. 6.3.2. Arguments for developing a new model To foster a more equitable environment for music research, a new publication model should prioritise inclusivity and accessibility. This means ensuring that scholars from diverse institutional and linguistic backgrounds are given the resources and platforms to publish and share their work, while creating space for the different types – COST Action CA 21161, Earlymuse 104 of material produced: academic texts, editions, digital copies, audio and video recordings as well as catalogues and database applications. Open-Access Publishing: One of the most promising ways to address inequalities in publishing is through open-access platforms. In the best case, open access provides a level playing field by removing financial barriers for researchers, enabling all scholars to share their work freely with the global community. In the context of musicology, this would allow valuable research on lessstudied repertoires, traditions, or sources to be disseminated without the financial burden that often accompanies traditional publishing. However, the question of funding open-access publication remains unsolved and perpetuates funding inequalities. Multilingual Publishing: To promote more equitable access to the scholarship, journals and publishers should embrace multilingual publication models. This could include bilingual or multilingual publications, along with translation services for non-English contributions. By offering research in multiple languages, scholars from around the world can engage with each other’s work, fostering a more diverse and international exchange of ideas. Examples are the already mentioned journals De musica disserenda and Anuario Musical, or series of critical hybrid editions such as Monumenta artis musicae Sloveniae or Corpus Musicae Ottomanicae or Monumenta Musicae in Polonia. Multimedia Publishing: To foster innovation in historical musicology, new publication venues should offer the technical infrastructure for presenting the wide range of research output produced in the field. Voix du Mercure Galant is an example for such an endeavour. The latest developments in artistic research and research creation are especially promising, but require technical support. Multimedia publication should be seen as an inspiring option, not as an additional complication. A growing trend in artistic research is hybrid publishing, which integrates multimedia content with analytical and reflective writing. For example, Music & Practice, and as mentioned earlier Anuario Musical, supports not only traditional articles but also contributions incorporating audio, video, and images. Similarly, platforms like ESMAE’s HTML-based model enable researchers to embed multimedia content directly into structured, navigable documents, without requiring coding skills. These models allow artistic processes and outcomes to be documented more faithfully than conventional print formats; moreover, – COST Action CA 21161, Earlymuse 105 by engaging with the material through multiple media, they potentially enable researchers to draw different kinds of conclusions than those typically reached through paper-based research alone. Collaborative, Scholar-Led Publishing Initiatives: A collaborative publishing network specific to musicology could be an important means of promoting equitable access to research platforms. By developing scholar-led, community-driven publishing models, such as collaborative online platforms or partnerships between institutions, music scholars could bypass the major commercial publishers. Metrics and Recognition: Traditional publication metrics, such as journal impact factors, often fail to capture the diverse range of musicological scholarship, particularly when it comes to research from less mainstream contexts or regions. A novel solution would incorporate the equality goals stated by the Declaration of Research Assessment (DORA). Several recent initiatives, exemplified by the case studies below, are setting a precedent for more inclusive practices. • The e-LAUTE project creates an open-access comprehensive and interactive hybrid edition of the lute tablatures of the Germanspeaking area between 1450 and 1550. • The research project Digital Presentation of the Long-SixteenthCentury Church Music Connected to Carniola provides much of its results (music editions as digitised post-editions, articles, chapters in monographs) in open access. • The Lost&Found Project at Lisbon NOVA University presents Medieval and Early Modern fragments of music as digital copies, editions containing reconstructions of missing parts as well as a database of analytical observations. • English Broadside Ballad Archive offers over 9000 ballads free to the public as text, art, and music. E-LAUTE browse view for 1540–1549. – COST Action CA 21161, Earlymuse 112 • Infrastructure: utilising existing networks and resources from both public and private entities. • Credibility: gaining trust and recognition from a wider audience through institutional backing. • Sustainability: creating a more stable and long-term model for the publication’s growth and success. It is undoubtedly illusory to imagine a single solution based on a concentration of actions, whether in terms of design, implementation, or maintenance. The risks of centralization are easily identifiable, being those of any monopolistic system. It would therefore be appropriate to deploy a shared solution based on the model of an infrastructure that would allow the combination of human resources, technical means, and scientific guarantees. The project-based research model proves to be an obstacle to long-term projection. Most musicological research laboratories do not have enough permanent staff and even operate in a hybrid model of perpetual and project-based resources, as is the case in several academies of science. This accumulation of pitfalls does not, however, prevent the development of a solution that would articulate the contribution of: • Willing research laboratories or university institutions, • Local or national facilities to house the data, • Private partners. By leveraging the strengths of both public and private institutions, this new publication model can achieve greater reach, impact, and innovation in the digital publishing landscape. Private partners are even more necessary because: • Some publication elements may come from private producers (recordings, films, editions) guaranteeing royalties to rights holders. • Communication methods are more solidly developed in certain private companies (this is particularly true in the music industries). There can be a dual presentation of musicological research: a first one responding to the new model and including all the elements of the research file (databases, catalogues, detailed descriptions, etc.); a second one manifesting in another format (a CD, a synthetic book, a radio program, etc.). And this dual presentation does not exclude other presentation modalities that could be imagined in the future. – COST Action CA 21161, Earlymuse 113 Public institutions like universities, research centres, and government agencies may provide stability and continuity in funding. This ensures the long-term viability of projects by enabling sustained research efforts without constant fundraising and by maintaining essential operations like server upkeep. Public funding also aligns with open-access principles, promoting societal impact by disseminating knowledge as a public good. Additionally, public institutions foster collaboration across universities, research centres, and international organizations—similar to strategic partnerships that drive innovation in the business world. To maximise the value of publications, a hybrid funding model that combines public and private support offers the most promise. Public institutions provide a stable foundation for long-term initiatives, while private entities contribute flexibility and additional resources. This collaboration balances risks and rewards, ensuring the sustainability and impact of scientific editions—much like strategic joint ventures in business. Encouraging private companies to sponsor scientific publications as part of their corporate social responsibility (CSR) initiatives can enhance their reputation while advancing knowledge— paralleling how businesses align operations with ESG (Environmental, Social, Governance) principles. Crowdfunding platforms also offer an opportunity to engage broader communities and stakeholders. This approach democratises funding while fostering a sense of shared ownership among supporters, creating a dedicated base invested in the success of scientific editions. Overall, developing viable business models for scientific editions requires a clear value proposition that emphasises societal benefits, educational impact, and innovation potential. Monetization strategies could include: subscription-based access to premium content, licensing fees for commercial use, specialised services like consulting or customised data analytics for private organizations 6.5. Long-term preservation Choosing an innovative form of publication, depending on the legislation of a given country, may mean that the publication will not be subject to mandatory protection and long-term preservation in its original form by the designated institution for the safe-guarding of national or local written heritage (e.g., a national library). In such a case, the creators must ensure and secure the publication’s durability themselves, as in many countries legislation has not yet responded to changing publishing practices and the digital dissemination of research results. – COST Action CA 21161, Earlymuse 114 Long-term preservation (LTP) of both data and software in new models for publication is something that has to be considered from the early stage of planning and creation. Over the years, heritage institutions have developed standards for LTP of digitised heritage objects, including paper-based scholarly writings and digital surrogates, such as e-books, e-scores and e-journals. New types of publications, that are not fixed and linear, or consist of various multimedia types, pose a challenge that need to be addressed. We already see digital humanities projects concluded some years ago, for which both data and software are no longer available. Both scholars and policy makers should think more about the long term perspective of maintenance and preservation before undertaking projects with such innovative digital products. The mere preservation of any digital data is not a technical problem, but preservation that enables the logical order of information to be recreated and managed in the future, in particular for continued inquiry, is. Archival systems dealing with preservation packages of data and metadata have their own implemented data models, designed for the needs of the objects they must manage. Bringing the products of digital multimedia musicology projects to one common denominator, one common data model, would be a huge challenge. Therefore, it is important to strengthen the knowledge of international standards and reference data models among musicologists so that their projects are compatible with standard data models (for example: IFLA LRM, CIDOC CRM, IIIF Presentation API data model, Web Annotation Data Model by W3C), thus facilitating the preservation of objects resulting from projects within existing systems, without the need to create new, one-of-a-kind, infrastructures, systems and processes. In any project dealing with data and creation of software, it is important to establish a strategy for long-term data storage, accessibility, maintenance, development, as well as the durability of the software necessary for processing, managing, presenting, and sharing data. Due to the human and hardware costs discussed above, as well as the necessary domain expertise input for maintenance, it may be beneficial and sustainable for musicology projects to partner with a stable public institution that specialises in long-term storage in accordance with international standards, such as those in the GLAM sector. – COST Action CA 21161, Earlymuse 115 6.6. Building public engagement into publication platforms Considering a wide spectrum of different musicological outputs as valuable for general knowledge can show new perspectives in the world of publishing as well. Integrating public engagement into scientific publication venues could feature tools and applications to: • Drive traffic to the platform (SEO, AI discoverability, linked content on high-traffic platforms to drive engagement with the publications), • Make research and its authors easily discoverable and digestible for producers of traditional media such as journalists, producers, media researchers, and for professional users in the cultural industries such as festivals, film companies, museums; and for policy makers, • Allow content to be embedded into other online platforms in different ways, given the CC Licence permits it, • Be optimised to integrate with social media if desirable, • Allow critical and creative re-use of content, • Allow audio and visual content to also be engaged with through sound systems and home cinema (e.g., audio playlists, screen casting), • Allow multiple pathways and views of a publication content so that different user groups can navigate it in different ways, • Facilitate engagement with specific interest groups, • Enable open community review and comment by communities of interest and practice outside of academia, • Facilitate participatory research (such as public contributing sources or annotations to a database), • Be compatible with accessibility standards, • Incorporate AI tools for searching, summarising and presenting content in different formats (such as visualisations, podcasts, blogs), • Gather data on users and usage, • Supports the inclusion of add-on features and widgets such as visualisation tools, analytical plug-ins. Such a platform would also need to be supported by a human infrastructure of professional skills and expertise that can enable researchers to get more engagement with their work. This could include training, consultation and practical support with: – COST Action CA 21161, Earlymuse 116 • Content creation and management for social media, • Media and public relations, • Making links with Culture and Creative Industries partners and relevant communities of interest, • Best practice in meeting accessibility needs, • Best practice in copyright and intellectual property, • Responding to and working with emerging technologies and standards, • Designing participatory features, • Working with programmers to develop features and add-ons, • Feeding back issues and suggestions for the future development of the platform. 7. Developing Communities Coordinating all these ways of knowing and representing music will require that we develop new communities of practice. Of course each scholarly discipline is just this: a community that has drawn up (with varying degrees of formality) some rules that govern evidence and how it is interpreted. What we would like to emphasise here is that a community of practice learns through community interaction.50 Now we need to imagine what sort of ‘communities’ might emerge as musicology learns to go digital, and what kinds of digital musicology might emerge from such collaboration. Learning how to use certain tools requires people to receive hands-on experience working on digital methods, converting them into digital practitioners. While digital specialists can provide this training, the project’s hosting institution also plays a role by ensuring that training, opportunities for practice, and collaborative spaces are available. Developing new communities of practice in digital matters requires an effort in building people, communities, institutions and their capacity. Here we consider a series of specific projects and deliverables in turn. 50 ‘Communities of practice are groups of people who share a concern or a passion for something they do and learn how to do it better as they interact regularly.’ Etienne and Beverly Wenger-Trayner (2015), Introduction to communities of practice: A brief overview of the concept and its uses. – COST Action CA 21161, Earlymuse 117 7.1. Building people We must not dictate the tools to be used. On the contrary, we should provide guidance on the available tools and train researchers to use the tool of their choice (the one they judge as the most appropriate for their purposes). This training can be done in different ways. • Training for Scholars, Editors, Archivists, Librarians, and Publishers, for instance through Summer Institutes (2-3 week programs) to build skills. Examples to consider: • Best Practices for Coding in Digital Humanities (Cambridge). • Cultural Heritage Data School (Cambridge). • DARIAH Training and Education (EU based and open, with a ‘registry‘ of offerings from Uppsala to Granada and from Dublin to Rhodes on many special topics, both as degree programs and stand-alone courses). • Training and education in imaging basics (DIAMM). • Edirom Summer School (Centre for Music – Edition – Media (ZenMEM) at Paderborn University) deals with research questions on digital editing and digital humanities, especially in workshops on TEI, MEI, and the Edirom tools. • Short-Term Scientific Missions (COST Actions), such as the ones organised by the present COST Action EarlyMuse CA21161. • Create Internships, Studentships, and Mid-Doctoral Fellowships. These can and do take a variety of forms at different institutions. Melbourne University in Australia, for instance, supports a Digital Studio program, in which students partner with faculty and technical experts to gain initial exposure to digital methods over about 45 hours to total work. A far more extensive form of applied digital skills training can be found at the University of Pennsylvania’s Price Lab for Digital Humanities. Here, a number of PhD students are offered mid-doctoral fellowships, again building on the idea that students can use these skills through their theses and throughout their careers. • In-person or online workshops on the use of a specific tool. Recordings of these workshops can be made and shared online to make the material globally accessible, addressing issues of inequality among institutions. • Creation of ‘Skills Network’ to find the connection among people, so that they can work on a peer basis. • Transforming professional development curriculum for musicologists and humanists. Enhancing the professional education with project planning and management, data literacy, data modelling, the basics of computer science, to facilitate communication and cooperation in digital humanities projects. – COST Action CA 21161, Earlymuse 118 7.2. Building institutions There are wide differences among organizations, and wide differences among their needs and capacities.51 Some institutions have invested heavily in the creation of specialist laboratories, including necessary digital research infrastructure and equipment. Centres for Digital Humanities, for example, have emerged recently as a model in many universities to serve as an umbrella institution for digital services, closely connected to those of the university library, and open to the entire community of faculty and students. In this way, digital projects in diverse disciplines can have access to the needed technology, under the guidance of experts to comply with standards. This model of shared infrastructure is equally valid for public institutions like libraries, which could serve the same purpose for the general public. Digital specialists in musicology, even if they do not easily fit in the traditional roles of faculty or staff, are clearly vital partners in all aspects of novel musicological digital publications, and would play an important role in any such Centre. Creating multi-year positions in such centres would also foster Citizen Science, innovative and sustainable digital initiatives, as well as collaboration among cultural and academic institutions. 51 See OPERAS-D’s ‘Developing network and e-infrastructure strategy. Report on online survey on optimizing e-infrastructure’ (2018) and OPERAS’s white paper on Best Practices (coming soon). Lodovico is an interactive, open, and transversal platform that collects the historical manuscript and photographic heritage in digitised form, from archives and libraries, with an openness to museum heritage. – COST Action CA 21161, Earlymuse 119 7.3. Building community Several projects and organizations have guidelines and handbooks for best practices which are extremely useful to build a community of practitioners. The Music Encoding Initiative (MEI) is a particularly good example of community organization, bringing together ‘technologists, librarians, historians, and theorists in a common effort to define best practices for representing a broad range of musical documents and structures.’ The results of discussions are formalised in the MEI schema, a core set of rules for recording physical and intellectual characteristics of music notation documents expressed as an eXtensible Markup Language (XML) schema. The MEI schema is complemented by the MEI Guidelines, which provide detailed explanations of the MEI model and suggested best practices. The schema is developed and maintained by the MEI Technical Team, an international group of experts. Other possible resources or models for Best Practices which promote knowledge exchange among people, therefore building communities of practice, are the following: • Handbook for Best Practices in creating and managing repositories (GitHub) • Best Practices for Digital Scholarship for authors, editors and publishers published by the Renaissance Society of America. • Awesome International Image Interoperability Framework (IIIF): Content includes helpful resources around each of the standard APIs, demonstrations of their use, tutorials and presentations. As they indicate, the list is especially helpful for orienting new community members and developers. • Best Practices Special Interest Group of OPERAS: Interest group of best practices in academic publishing. Among the technological topics discussed by the communities are the following: • Digitisation and handling of special collections. Some information about this can be found in FADGI Technical Guidelines for Digitizing Cultural Heritage Materials (3rd Edition, 2023), DFG’s Practical Guidelines on Digitisation (2009) and DIAMM’s Technical Overview website. • Documentation and version control: Using version control for documenting workflows and practices followed by the institution (e.g., GitHub: https://docs.github.com/en/get-started/using-git/about-git). • Workflow management: Determining the steps and tools needed to take the source to the screen, how to keep track of the completed steps and work of the multiple participants involved. • Data about music: Metadata (who, what, when, where), using Authority Files and Identifiers (VIAF, GeoNames, and so on). – COST Action CA 21161, Earlymuse 120 • Work authority registries. • Role authority registries for performing arts. • Ontologies for Music and Intangible Heritage more generally. • Relational Databases and how to query them. • APIs (for human and machine discovery). • Reusable Data: Flat CSV or JSON files as ways of preserving data beyond the lifespan of any specific application that might deliver it. • Data Cleaning: This is a major part of every project, especially since many projects have long histories as a series of informal spreadsheets or other legacy work. • Data Validation: Process of verification that your data is well formed, and learning how to update your data to conform with current practices and schemas (e.g., how to use XSLT to automatically update your XML or MEI file to the current schema of the encoding). • Music as data: Encoding Scores, Recordings. • Retrieval and transformation of information: For example, transformation from one encoding format to another, transformation from an encoding format to PDF or to audio, retrieval of the parts from a MEI score, or the metadata in and out of a MEI file, scraping data from a website. • Data visualization: Systems like Scalar (an open-source system for publication using rich multimedia), Verovio (an open source tool for rendering high-quality, digitally engraved music scores), and Plotly (an open source Python library for graphs and charts). • Identifying missing or corrupted data: Ways to facilitate the identification of gaps in data, how to automate this, how to clean/repair data. • Summary of digital tools for specific purposes of manipulating or generating music data: Tools for optical music recognition (for common Western music notation or for early music notation), tools for data analysis of music (e.g., music21, Humdrum, and CRIM intervals), editors for encoded early music notations (CMME editor, Measuring Polyphony Editor), editors and renderers for encoded music (e.g., mei-friend and Verovio), and others.52 • The Quotable Text: IIIF (with support for images, audio, and video), Quoting Digital Scores (EMA). 52 See Laurent Pugin, Martha Thomae, Marnix van Berchum, Kévin Roger, and David Fiala (2025), ‘A new XML conversion process for mensural music encoding: CMME_to_MEI’. – COST Action CA 21161, Earlymuse 121 • Annotation (related to previous). What can you ‘add’ to a digital text? • Leveraging machine learning to help build or query data. • Rights Management (how and when you can reuse things). Implementation of a process for converting CMME files to MEI. 7.4. Building Capacity As musicology progresses into the ever-more-digital future, there is a need for sustained and sustainable efforts in tool development (also consult OPERA’s white paper on Tools Research and Development). Here we note both desiderata and currently-available tools that are promising for future development, in order to provide inspiration: • Last Kilometer Tools that will help things move from source to screen. – Editorial tools for creating digital scores (such as MuseScore plug in, or enhancements to MEI Friend). – Viewers that would take encoded scores and put them online, either standalone or as part of a Digital Musical Example engine (e.g., using Verovio). These could be on some kind of free license for publishers to use. Here, for instance, is an example of a REST interface that returns MEI encoded examples to the browser, with highlighted elements to represent analytic annotation; code at: https://github.com/eleon024/ema_ – COST Action CA 21161, Earlymuse 128 early recordings, is inviting the general public alongside academics to contribute to the database and publications. The UK’s National Coordinating Centre for Public Engagement describes public engagement as ‘build[ing] an inclusive higher education sector where communities can contribute to, and benefit from, knowledge, teaching and research.’61 The Centre emphasises active public participation and contribution, rather than passive consumption, to enhance the quality, relevance and accountability of publicly funded research. At the EU level, Engage2020 focused on mapping and exploring public engagement methods and tools to make European research more inclusive. The project produced resources detailing current practices of public engagement, aiming to improve the two-way communication between scientists, policymakers, and citizens to ensure regular exchanges of views. The Knowledge Valorisation Platform undertook analysis of citizen engagement for knowledge valorisation practices, drawing on 60 selected case studies from 37 countries, across the EU and internationally, mapping the benefits of participatory processes and exploring key elements for successful valorisation of knowledge and research results with the engagement of citizens. The Mutual Learning Exercise (MLE) ‘Public Engagement in R&I’ (2023–4) aimed at exchanging success stories, experiences, and challenges in fostering meaningful public participation in research and innovation. However, none of the reports produced by these initiatives contain any reference to music or musicology. This highlights the need to consider public engagement more fully in relation to musicology in Europe. There is a need for a focal point to foster discussion and share best practices in this area, in order to develop a shared understanding of how musicology can respond to the wider imperative for public engagement. 8.3. Research Policy and Public Engagement Horizon Europe places significant emphasis on citizen engagement to align research outcomes with societal needs, through meaningful engagement with and integration of citizens’ needs and interests into the agendas and outputs of publicly funded research. This focus is articulated through 61 NCCPE (2025), ‘Homepage’. – COST Action CA 21161, Earlymuse 129 the emphasis on citizen science and dedicated funding for public engagement. However, the design of these initiatives is largely concerned with research in the sciences, while the Creative Europe programme deals with public access to the arts. The focus of Creative Europe is on the Cultural and Creative Industries,62 rather than academia, hence, the program is not geared towards musicology. Case Study — The UK Research Excellence Framework (REF) The UK context offers an example of how changes in research policy can be achieved to foster public engagement with the field of musicology. The REF (the national research assessment exercise) defines ‘Impact’ as ‘an effect on, change or benefit to the economy, society, culture, public policy or services, health, the environment, or quality of life, beyond academia.’63 Impact results from Knowledge Exchange, which is defined as a ‘collaborative, creative endeavour that translates knowledge and research into impact in society and the economy.’64 In previous REF submissions, subject areas were required to submit ‘Impact Case Studies’ whose weighting in terms of funding return were several times that of an academic monograph. This led to impact being embedded as a strategic area for development and support in UK universities, with specific resources and funding streams available to support the design and measurement of impact activities. However, with a narrow definition of impact on easily quantifiable effects and changes to wider society in the UK, REF often excluded the kinds of public engagement activities that music and the arts more generally excel in, making the wider value of the research difficult to report. This meant that these subject areas and individuals did not always receive due credit for them and such activities were not necessarily supported by institutions or rewarded by the REF. Broadening the scope of assessment to incorporate Public Engagement alongside Impact is expected to have a material effect on Public Engagement strategies and support within institutions, as well as the credit given to the academics who excel in this area, stimulating more activity in this area. A robust funding structure that covers the technical costs of innovative dissemination formats but also the work hours for public engagement that are not usually specified in academic employment contracts is obviously required as the basis for these developments. 62 European Commission (2025), ‘Music - Culture and Creativity’. 63 UKRI (2025), ‘How Research England supports research excellence’. 64 As formulated in the Knowledge Exchange Concordat. – COST Action CA 21161, Earlymuse 130 8.4. Considering impact strategies in musicological research projects65 The notion of impact refers to the deep consequences of academic research on the general public, outside the scientific community. Impact can be cultural, economic, environmental, social, healthrelated, legal, or even technological. In the United Kingdom, it represents a key issue for university funding, meticulously documented for submission to the Research Excellence Framework66 and demanding long-term strategic planning. In Continental Europe, this concept is beginning to take hold, with more national research funding bodies adopting criteria related to societal impact that move beyond scientific valorization of published research results. It is crucial to take impact into account in all stages of scientific projects, from initial planning to research, analysis and outputs. Impact can take place during the funded phase of a project, but frequently develops after the research team has dispersed: this creates a positive legacy, but also poses challenges for the management and measuring of impact. In the UK, the underpinning research of the project also has to be published before impact can be said to derive from it: another challenge to be met. The experience of recent collaborative research demonstrates that devising and delivering impact can encourage unexpected research paths: the impact deriving from one pathway, but also stimulating subsequent and unanticipated strands of research. Consequently impact can allow us to imagine new ways of approaching the discipline, as well as sharing results with a wider audience with long-term influence on the perception of music history. Over and above the academic evaluations that are necessary to obtain the funding inherent to the continuation of research, taking impact into account is invaluable for long-term dissemination. Impact allows us to think about how we can share the results of our research and rethink our publishing practices beyond our usual ecosystem. How do we visualise and communicate rich data, or link musical data and history? What are our practices in investigating the specificities of a musical period? Finally, how can we actively engage the audience in crafting the research beyond standard period dissemination? 65 This section is largely inspired by Braud et al. (2025), ‘Considering impact strategies in historical and musicological research projects’. 66 For more information, see UKRI (2025), ‘How Research England supports research excellence’. – COST Action CA 21161, Earlymuse 131 Thinking about impact from the inception of a project allows us to consider the global scope of scientific spin-offs, integrating the wide sharing of knowledge into the very heart of the scientific proposal. Taking this a step further, it is thus possible to use societal observations linked to the stateof-the-art of a discipline as a starting point for questioning it, in a sort of reverse engineering process made possible by reflecting on the importance of impact. Implementing impact strategies must aspire to unite communities around perennial scientific themes; it is therefore important that events organised as part of a project gather a wide audience in contrast to a niche of researchers. In fact, we should specifically commit to targeting audiences peripheral to the world of research in order to demonstrate inclusivity in changing the understanding of historical and/or musical facts in a thorough way. Among numerous impact cases studies, the notion of identity is regularly mentioned. It is a focal point around which we can build projects, both from the point of view of impact and of the research itself. We can thus question the nature of identity in the context of the study of music, where the coalescence of many parameters shapes the political and historiographical role of a work, as well the identity — both as essence and as performability — of its performers, composer and the people to whom it is tied. By attempting to recreate the performative contexts in an immersive way, it seems possible more thoroughly to connect the identities being studied with the identities of the audience. Even more accessible to a range of communities (for instance, people’s homes) than live performance is the video medium, whether through films or documentaries. However, working on this kind of media comes with challenges such as very high production costs and difficulties in translation (as discussed below). In order to give an account of the scientific work carried out as part of a research project, a process of translation needs to be considered to demonstrate the depth of the issues studied. This rendering process does not only target the purpose of the project in terms of impact — the communities identified during the prospects, that will be met during later events — but also the various agents involved in the logistical organisation of events, external partners such as musicians and singers, sound technicians, etc., in short, all those who make the impact possible by internally shaping it. Mutual understanding is thus a key resource for healthy exchanges and helps to give a new impulse to the discipline, from – COST Action CA 21161, Earlymuse 132 the inside. One should also be wary not to confuse this work with cultural mediation, which dwells in explaining artistic practices through example and talks. Despite the efforts made, the vast majority of researchers report that there is still a palpable resistance from the public to musical heritage, whereas patrimonial and historiographical elements seem to be more easily perceived and integrated into existing knowledge. Although this can be explained by a tendency to feel more at ease with historical figures and their identities than with musicians, even less than with works and their contexts, this still shows the reach of the work needed to make a profound and authentic impact on diverse communities throughout the presentation of musical works. Moreover, unlike its historical counterpart, musicology often tends to be viewed negatively by the scientific community, particularly in terms of its ability to legitimise an impact on the public. For example, rejections of scientific projects have sometimes been justified in reports by a simple lack of interest of the reviewers in 16th-century music, without any other explanation; this can be a major problem among funders with multi-disciplinary peer reviewers. In a context where funding is becoming increasingly competitive, these situations lead to considering multidisciplinary projects in order to integrate the musical experience into social, political and historiographical frameworks. In this respect, the example of the ‘Henry VIII on Tour: Landscapes. Communities and Performance’ project led by Anthony Musson with John Cooper, Magnus Williamson, and Kate Giles, is worth considering. Their impact strategy is structured around four main areas unified by the reconstruction of Henry VIII’s geographical journeys: • Raising awareness for local communities of the existence of places crossed by Henry VIII through site visits. • Archaeology of pre-identified sites: discovery of the past combined with community sharing open to all ages, enhanced by digital visualisations of the site. • Reconstruction of ceremonies of progress including actors and reproducing banquets: inclusion of communities in the stories of Henry VIII’s hosting through convivial and festive moments. • Musical performances of newly edited music, in accordance with the corresponding liturgy, given in preidentified places of progress: participants and communities experience service enactments. While some of these suggestions may seem very far away from the European researchers’ daily practice, they constitute a key standard for the UK practice of impact and show us that interdisciplinarity and collaboration between researchers and a wide range of seasoned professionals are absolutely – COST Action CA 21161, Earlymuse 133 fundamental to interfacing with communities. It is thus necessary to plan for these tasks as soon as the project proposal is submitted, since they require a high level of financial and inter-personal investment, as well as carrying out preparatory surveys prior to submitting funding applications. Those will enable us to identify the key operators with whom to collaborate: associations, theatre companies, archaeological centres, but even ecclesiastical communities or caterers, as seen in Henry VIII on Tour’s banquets! Having already strong links tied with the communities, they have much to offer to us in terms of better understanding of impact. Considering the audience is one thing, but it is also important to think about how research projects fit into the communities’ environments. Drawing on historiographical data is a first step towards rediscovering heritage sites, as those places can hold significant meaning for individuals, highlighting the connection between the appreciation of places and mental health. In this regard, music seems to be a perfect candidate for crafting the sonic ecology of a place through imaginative use of sound and space through a real musicological expertise. Spotlight: ‘Listening to the Commons’ This project is testimony to how impact can stimulate new research ideas rather than dictating the research, and allows the building of further links between the research fellows, unexplored places and new communities. A project team combining historical, acoustic, and digital humanities expertise succeeded in showing that, thanks to a ventilator driven through the old House of Commons ceiling, women (formally excluded from political life) could access debates in the Commons with in some cases a better listening experience than the male MPs down below. The study covers the period from 1820 to 1834, taking the overall chronology of the project to six centuries. Twenty-five tickets were available in the spaces accessed through the ventilation ducts, and it was possible to buy drinks, making it a social space. Two sketches of the space by women survive. In 2018, the Voice & Vote exhibition in Westminster Hall (107,000 visitors) recreated a replica of the ventilator, simulating how women may have experienced the sound of the House of Commons. The exhibition was curated by senior parliamentary staff, but dialogue with the research project enabled them to upscale the visitor experience with ‘authentic’ sound: one example of the dynamic relationship between host institutions and research-derived impact. A historic debate on the slave trade was also staged within the Commons chamber, drawing on project expertise and calling on Members of Parliament and parliamentary staff as ‘actors’. Two concerts of the sacred music of Nicholas Ludford were also performed, drawing on the Parliamentary choir and the choir of Gonville and Caius College Cambridge, in spaces associated with St Stephen’s Chapel: putting music probably composed for pre-Reformation St Stephen’s back into its original space. Events were filmed, and also audio-recorded for use in future acoustic modelling. These events wove a strong connection with two communities: visitors to the Palace of Westminster, and the politicians and staff who work there. Feedback forms showed that the exhibition played a role in profoundly changing attitudes, as well as inspiring new stories and ideas. – COST Action CA 21161, Earlymuse 134 Spotlights: ‘A Musical Journey’ at the KBR Brussels The KBR Museum in the Royal Library of Belgium has recently opened the exhibition A Musical Journey, centered around the library’s rich collection of illuminated manuscripts of Franco-Flemish vocal polyphony from the library of the Dukes of Burgundy. The manuscripts are displayed with other artworks and objects from the period. Characters from the manuscript illuminations are projected on the ceilings, and visitors can listen to music and stories at audio stations. In addition, the exhibition is framed by concerts and workshops, and the library blog features a scholarly but generally accessible article by Brett Kostrzewski from KU Leuven on the musical repertoire. While being highly successful manifestations in terms of impact, it must be stressed that the human and logistical cost of these events is colossal. Many events require legal authorization or the build-up of trust over many months, not to mention the staff mobilised as research support. It would therefore seem that the quality of impact obeys a circular logic that offers the bestendowed establishments the opportunity to organise more efficient impact strategies. Although music researchers are amplifying their efforts with endeavours to promote the results of scientific research to the public, they still inherently face resistance from funding institutions. When applications are submitted, it is not uncommon to come across remarks such as ‘what’s the point of having musical performance ?’... It is therefore up to the researcher to rethink the term in order to give it an interdisciplinary value — by referencing the performance of ‘experimental archaeology’, for example. In an era where sharing music has been completely rethought through the internet and streaming in particular, it seems more necessary than ever not to miss the step to connect to the cultural industries, so as not to become even more isolated than we already are. Collaborating with the art-de-vivre industry as well as fashion may seem far removed from our concerns; these collaborations do, however, make sense in the context of refined reconstructions of musical performances, bringing together newly rediscovered or edited music to all sorts of communities and identities interested in clothes. Although this might be truer in the UK, where history courses offer a very visual approach to the Renaissance, we still have many lanes to explore in order to offer immersive experiences that can raise the awareness of diverse communities. In conclusion, musicology is equipped to take on the tasks that are crucial for accessing, safeguarding, and disseminating cultural heritage on the basis of scientific methodology and standards. The visibility and accessibility of the discipline in society at large must however be increased in a way – COST Action CA 21161, Earlymuse 135 that musicologists are perceived as the experts they are. However, both infrastructure and funding options must be improved. 9. Communication and collaboration Music plays a vital role in shaping cultural heritage, identity and memory, but it has been underrepresented in cultural heritage initiatives. One reason for this could be its hybrid nature. It is both tangible in manuscripts, prints, instruments, paintings, etc., as well as intangible in the nature of performance and the knowledge required to interpret and bring sources to life. While this duality has often been flattened out in cultural heritage initiatives, the way forward is to celebrate it. The experience of people working within the broader umbrella of cultural heritage shows how important it is to approach heritage not only as a static legacy of the past but as something dynamic, constantly reinterpreted and renegotiated in the present. At its core, music, and the music scholarship that invigorates it, is the apotheosis of living heritage, a shared resource that shapes collective identity and memory. Musicology needs to move beyond seeing itself as a niche academic field to claim its rightful place as a torchbearer that transforms scholarship into a living practice: attending to the past while looking toward the future. There are many possible paths to achieve this end, but they all have a few elements in common. People working in musical scholarship need to come to a consensus on the best ways to work together towards common goals. They need to reach outside of music history circles and engage with policy makers, funding organisations and culture heritage promoters across disciplines. In policy and legal discussions, it is crucial to make a compelling case for the richness and distinct needs of music heritage, highlighting how it differs from other heritage domains in its reliance on sound recordings, performance practice, and complex metadata. In short, we need to shift the narrative and take our place as leaders in the cultural heritage space. People who work under the broader umbrella of cultural heritage — especially those who receive research grants aimed at preservation, interpretation, and dissemination — can teach us a great deal when it comes to the heritage of music. Musicologists should learn from them, and adopt their methods of community engagement and public history projects. This would significantly influence the direction of policies and research and cultural programs. – COST Action CA 21161, Earlymuse 136 Musicological research needs to highlight the ways in which it combines traditional research methods with digital innovation. By embracing both analogue rigor and digital innovation, we can rethink what it means to do musicological research today, especially through hybrid models which combine traditional research methods with digital innovation in interdisciplinary teams. These kinds of collaborations could provide a realistic pathway towards embedding musicological research in European digital initiatives and could prove successful in acquiring funding. * * * We conclude this document with several actions that we musicologists can already undertake, focusing first on legal reforms and policy and then on project-based actions that must follow. We provide examples of potential partners. → Active participation in consultation processes, working groups, engaging with policymakers at the European and national level. Active participation means, for example, drafting Position Papers outlining why music should be a priority in EU heritage policies or about the unique value of music heritage, requesting meetings with relevant EU cultural policymakers, or maintaining a seat at the table during critical legislative discussions such as the upcoming review of the EU Copyright Directive in 2026. Potential partners include: • DG EAC (Directorate-General for Education and Culture) • DG RI (Directorate-General for Research and Innovation) • DG CONNECT – Directorate-General for Communications Networks, Content and Technology • Committee on Culture and Education • Council of Ministers for Culture (to ensure national-level buy-in) • ECCCH (European Cloud for Cultural Heritage) The upcoming 2026 evaluation of the Copyright Directive as highlighted by Henna Virkkunen, Executive Vice-President for Tech Sovereignty, Security, and Democracy, presents a critical moment – COST Action CA 21161, Earlymuse 137 for stakeholders to advocate for reforms.67 The time is now, then, to argue for solutions that better accommodate musicological research, digital preservation, and metadata accessibility. As AIgenerated content, digital musicology, and text-and-data mining (TDM) become increasingly relevant, the upcoming review offers a strategic opportunity to push for legal adaptations that ensure fair access to historical recordings and metadata for research purposes. → Advocating for music heritage-specific exemptions in copyright law. This could include explicit legal exemptions for heritage institutions, clarification of metadata usage rights and greater alignment with fair use models which recognises that research-based exploitation of copyrighted works does not necessarily undermine rights holders’ financial interests. We recommend using the Fair MusE Deliverable reports and European Heritage Hub Policy Monitor as examples and foundations. As the EU considers updates to its copyright framework, it is imperative that the voices of musicologists, heritage professionals, and researchers be included in these discussions. By positioning musical heritage as a key pillar of digital innovation and European cultural policy, we can help shape a futureoriented legal framework that not only protects intellectual property but also fosters research, accessibility, and preservation in the digital age. → Promoting music as cultural heritage in the EU engaging with institutions, NGOs, and lobbying organizations that focus on culture, heritage, and creative industries. The following are some potential partners: • Creative Europe: The EU’s flagship program supporting the cultural and creative sectors, including funding for music as heritage. • European Heritage Label (EHL): Recognises cultural heritage sites, and music traditions could be integrated into this framework. • Music Moves Europe (MME) • Europa Nostra: A major pan-European NGO dedicated to cultural heritage advocacy. 67 European Parliament (2025), ‘Commitments made at the confirmation hearings of the Commissioners-designate 2024-2029’.