Introduction

The digitisation and/or digitalisation of library catalogues has been one of the most important developments in the recent history of the library sciences, having a profound impact on the ways in which users locate and access a host of scholarly resources including manuscripts (Chambers 2013; Earnshaw and Vince 2008; Bhardwaj 2018). Prior to the advent of the digital age, library catalogue records were largely static; whilst new cards could be added to a card catalogue to reflect additions to a collection, the entries themselves were rarely revised. The confines of space imposed by these card catalogues further inhibited cataloguing practices, reducing complex archival materials down into neat summaries, often of a purely bibliographical nature. The same was true of printed catalogues. Whilst these could, and often did, follow descriptive cataloguing practices to provide more than mere bibliographical data, the entries were made static by their printed form; and the requirement to produce a manageably-sized book restrained the levels of description that could be included. Additions could also only be made by publishing a revised edition of the catalogue, often a costly and time-consuming venture. Thus, whilst these catalogues provided an invaluable service by recording vast collections, they were often inadequately designed for assisting researchers tackling complex research questions, being by necessity brief and typically bibliographically focused.

Beginning in 1971 with the launch of the University of Ohio’s computerised catalogue, however, digital tools have been applied to these previously printed catalogues, providing opportunities for descriptive cataloguing to take precedence and for catalogues to better represent the dynamism of the items they are cataloguing (Miller 1980; Nilges 2006). An important first step in the transformation from paper to data was to OCR existing printed catalogues and link them to machine-readable cataloguing (MARC) authority files (Tseng 2001; Scott and Vinson 2017; Bowman 2006; Harrison 1985; Allen 1987). This initial step towards transforming print into metadata would later enable cataloguers to encode these catalogue entries using markup languages such as TEI/XML (Nellhaus 2001; McDonough 2009). This type of encoding has two key benefits. First, the creation of metadata allows cataloguers to establish linked entities within their collections, enabling all items relating to a single individual, for instance, to be linked together in a subset of the data. In addition, the widespread use of such encoding across institutions ensures interoperability, enabling individual library systems to communicate with one another (Allen, Kulczak and Gilbertson 2017). The benefits of this have been well established with the creation of vast union catalogues such as WorldCat or ArchiveGrid that actively use linked data to enable researchers to search for items across thousands of repositories (Jordan 2011; Orcutt and Powell 2006; Sneary 2006).

It is now considered atypical of institutions to not offer a digital catalogue of some kind as the digital has come to be an expected norm given its ability to improve access to library collections. However, many online catalogues fall short of their potential. Whilst using OCR to transform print catalogues into digital ones has made these catalogues more accessible, many institutions have stopped there, being inhibited by a lack of resources. As such, many digital catalogues perpetuate the brevity of their printed counterparts despite no longer being constrained by space or the inability to swiftly make changes to a record. In doing so, such catalogues also limit user interactions and discoverability, lacking the levels of detail necessary to support complex research questions. Additionally, the bibliographical entries typical of card or print catalogues often privileged outmoded, biased and problematic perspectives (Drabinski 2013; Moorcroft 1993; Littletree and Metoyer 2015). Whilst research trends have continued to make great strides, radically altering scholarship and reflecting critically on past biases, catalogues have often failed to keep pace despite their digitisation, reflecting the values of the past whilst adopting a new digital format.

Beginning with the premise that cataloguing practices must continue to evolve in order for digital tools to better support researchers and aid new discoveries, the Catalogue of Manuscripts Relating to Voltaire/Catalogue des manuscrits relatifs à Voltaire (CMV)¹ has actively sought to push the boundaries of what constitutes an online catalogue, building a digital tool that will not only enrich the research experience and allow for previously obscured connections to be drawn between manuscript materials, but that also sets a precedent for future catalogues by demonstrating new ways in which these tools can better support current and future research trends. Just as the materials contained within a catalogue are nuanced and complex, so too must the catalogue itself be malleable and dynamic, working to enhance access to archival collections and to support researchers rather than inhibiting them.

What is CMV?

CMV is a digital union catalogue that offers detailed, granular descriptions of manuscripts concerning Voltaire that are preserved in collections around the world, bringing these records together for the first time in a single resource. Work began on the project in 2022, with generous support provided by the Astra Foundation allowing the creation of an open-access resource that enables sophisticated searching. Its architecture has been built with intra- and interoperability in mind, allowing fresh connections between documents to be drawn through its levels of descriptions, which will enable important benefits for scholarship on Voltaire and the Enlightenment.

The corpus of manuscripts relating to Voltaire is vast and heterogeneous, and we are only beginning to understand the size and shape of the collection thanks to the construction of CMV. Although manuscripts produced by Voltaire himself, or on his behalf by his secretaries, constitute the very heart of this catalogue, it is important to remember that Voltaire’s image was shaped by the emergent cult of celebrity in the eighteenth century (Lilti 2014; Goodman 2016). As such, his manuscripts were continually copied and recopied over the course of the Enlightenment and beyond, and indeed were reproduced and edited for a variety of different reasons and in a diverse array of formats. These copies are a useful means through which to gauge the importance of Voltaire’s manuscripts and their reception through the ages, and thus bring a significant new dimension to our understanding of Voltaire, alongside the originals. CMV therefore takes the manuscript culture of Voltaire in its broadest sense, considering it as not restricted to his own writings, but as including their various subsequent permutations. By adopting this broad approach, the catalogue allows users to consider both Voltaire’s own manuscript practices and his wider reception in 18th-century culture and beyond, offering a more rounded insight into Voltaire and manuscript cultures of the period.

The catalogue will contain approximately 20 000–30 000 entries for manuscripts produced by, or concerning, Voltaire. The source base it draws upon is thus both rich and diverse, including his abundant correspondence, drafts and copies of his works, diary entries by others giving an account of meetings with him, and even handwritten notes at the end of an 18th-century Jamaican almanac that quote from his works. Whilst the sources themselves are diverse in their nature, they also come from a wide array of collections scattered across the globe. Major collections held at the Bibliothèque nationale de France (BnF), the Morgan Library, the National Library of Russia and of course the Voltaire Foundation are all represented in the catalogue. However, it also features items held in smaller and/or lesser-known collections such as Waseda University Library in Japan, the University of Cape Town’s special collections, and a host of local or provincial archives such as Hertfordshire Archives and Local Studies Centre. Additionally, by recording items held in private collections (notwithstanding the sometimes more limited information that it is possible to display about these), CMV attempts to capture those more transitory manuscripts that appear in public view only fleetingly. CMV is ambitious in its scope and seeks to record all of the known manuscripts relating to Voltaire at this moment in time. However, given the digital nature of the resource, CMV can also accommodate new discoveries as previously unknown manuscripts emerge. As will be discussed later, the back end structure of the catalogue allows newly added data to appear on the front end rapidly, with new or updated records being able to go live for users in a matter of moments. Freed from the confines of print, the resource can thus continue to offer users the fullest reckoning of manuscripts relating to Voltaire in perpetuity, ensuring the value of the catalogue to researchers for many years to come.

The ability to rapidly update sources, however, has benefits that reach beyond our ability to provide the most complete record of manuscripts relating to Voltaire. In recent years, it has been acknowledged that catalogues often perpetuate and/or create biases through their use of language and the hierarchies within which subject terms are used. Movements such as CritCat advocate for a more critical lens to be applied to cataloguing practices in which these biases may be acknowledged, challenged and removed, either through the adoption of alternate vocabularies and structures or through the inclusion of content warnings.² Whilst CMV has followed CritCat guidelines in an attempt to be as inclusive as possible, language is an ever-changing phenomenon and we recognise that terms that are widely used today may be reflected on more critically in the future. The ability to revise entries thus allows us to continue to ensure that the catalogue not only represents the materials of the past, but also reflects the values of the present, creating a resource that is as inclusive as possible and that continues to reflect critically on its own practices.

In its depth and granularity, CMV pushes the boundaries of what a manuscript catalogue can achieve, going far beyond the typically bibliographic entries associated with summary cataloguing, and extending the content usually associated with descriptive cataloguing. Where possible, entries record not only standard catalogue data but also often overlooked, but nevertheless crucially important, elements of a manuscript such as marginalia, watermarks and signs of use, innovatively allowing users to search and filter the database based on these criteria. In doing so, the catalogue addresses a variety of research needs, from potential ambitious ‘big data’ visualisations to the most cutting-edge studies of the materiality of the text. Thus, the novel depth of CMV enhances the user experience by offering fresh insights and access to previously obscured data in order to generate new discoveries and support current and future research trends.

How has CMV been constructed?

As outlined above, the materials included within CMV are diverse in their nature, location and levels of previous cataloguing. It has therefore been necessary to build a union catalogue that can handle these differences, imposing a standardised framework that is flexible enough to account for the unique nature of each source whilst maintaining the best possible user experience. This means that catalogue entries are not simply copied and pasted from their institutional catalogues, but have instead been revised, and in many cases expanded, to ensure that every entry complies with both International Archival Description Standard and the standardised typologies adopted by CMV to improve searchability and maintain consistency throughout. By adopting such an approach, we have been able to ensure cohesion across the resource and whilst some records are by necessity currently limited, either due to lack of access to the manuscripts themselves or because they have previously only been catalogued in a very limited way, the ability to continually update and enhance records will allow us to improve upon these going forward.

The fields included, as discussed above, far exceed those of existing catalogues in their depth and scope, offering detailed descriptions of both the contents and materiality of the manuscripts they record. Table 1 provides a summary of the 43 fields employed within CMV, offering a sense of the depth and breadth of each manuscript record.

Creating an online resource that could accommodate this number of fields in a useable way required more than just technical knowledge: it demanded a blend of creativity and strategic thinking. The project started with a clear objective: to create a database that not only lists manuscripts but also allows researchers to explore them in an engaging and meaningful way. To achieve this, a hybrid approach using PowerBI and WordPress was chosen.

PowerBI is known for its robust data visualisation capabilities and allowed us to create an interactive map of the data, dubbed the ‘Discovery tree’, where users can discover manuscripts based on various attributes such as origin, author, genre and keywords. The visual nature of PowerBI means that this data can be browsed dynamically, allowing the user to explore a complex dataset visually, rather than through static lists. Such visualisations will be of especial use to those who are new to archival research, such as students, enabling them to navigate through the data in an accessible and engaging way.

WordPress, on the other hand, has provided the foundations of a user-friendly interface where researchers can conduct advanced searches and browse data more organically. WordPress’s flexibility crucially allowed for customisations that have been essential in meeting the complex and unique needs of CMV, as will be addressed in more depth below. By combining these two platforms, CMV harnesses PowerBI’s visual strengths alongside WordPress’s user-friendly data-management capabilities, resulting in a resource that is able to support an extensive and complex dataset and to present this in an accessible and engaging way for users.

Creating a manuscript catalogue is no small feat, especially when aiming to present data in a manner that allows for intra- and interoperability. One of the key challenges faced in the development of CMV was taxonomies and the handling of the multitudes of attributes each manuscript can have. While WordPress supports custom taxonomies, it does not natively accommodate sub-attributes for these taxonomies, requiring the development of custom plugins to manage these needs. For instance, each manuscript is located in a specific repository that itself has additional data in the catalogue such as the location of the repository, its International Organization for Standardization (ISO) and MARC codes, and contact details for the special collections department. Custom plugins have enabled CMV to manage these additional needs, ensuring that every piece of data, including sub-attributes, can be accurately represented and easily accessed.

The second challenge presented by taxonomies was how to make them visible and explorable on the front end. Each manuscript record features a series of attributes, such as people, origins, repositories and keywords that we wanted to be clickable, featuring a corresponding landing page that displays all records sharing the given attribute. If, for instance, users were to view a record written in the hand of Voltaire’s secretary, Jean-Louis Wagnière, we wanted them to be able to see all of the records within the catalogue of items to which Wagnière was a contributor. To achieve this, custom templates have been created so that each primary attribute has its own corresponding landing page that displays its unique set of data. These templates have made it possible to achieve a browsable list of related manuscript records, accessible from any given record through a clickable link, making the user experience more intuitive and enabling users to easily identify related items within the broader dataset.

Given the complexity of the attributes and their multiple interrelations, the architecture of the database had to be meticulously planned and implemented. This was especially crucial when considering the usability of the catalogue, as it soon became apparent that searching for data and exploring data are two very different matters. Search boxes, both advanced and simple, are ideal when researchers know what it is they are looking for; for others, browsing a dataset gives the best results. As such, the online resource needed the functionality to facilitate both searching and exploring.

One of the more innovative aspects of CMV is the use of visualisations typically reserved for financial and/or sales data, namely the decomposition tree, renamed the ‘Discovery tree’ as noted above. This tool (Figure 1) allows researchers to explore data by various attributes, with the ability to adjust them dynamically to show related data with each click. For instance, users, through a selection of filters chosen by themselves in their preferred order, may view all records for manuscripts written by Voltaire in London within the genre of poetry that are classified as presentation copies. Should they wish to, they can dynamically change the order of attributes by dragging and dropping them into their preferred order to create a unique pattern of results that best fits their needs. Applied attribute filters can also be added and removed at any stage, again allowing the user to customise the visualisation.

Figure 1

The Discovery Tree

The development of CMV has thus highlighted the importance of both making high-quality data searchable and discoverable, and the usefulness of using detailed data recording to provide the basis for innovative and engaging visualisations. The creation of CMV was not only about developing a tool, but also about pushing the boundaries of what digital catalogues can achieve and thereby transforming the ways in which researchers interact with historical data. By combining the visual strengths of PowerBI with the flexibility of WordPress – especially as it is enhanced by custom plugins – we have been able to create a resource that is innovative, powerful and, crucially, user-friendly.

Impact and future use

CMV thus boasts a variety of facets that have proven that it is possible to significantly enhance digital catalogues in order to best support diverse users and their research questions. In order to demonstrate the impact that this resource will have on scholarship, we would like to highlight a few areas in which these developments will be of great value to the scholarly community.

The linked attributes used throughout CMV enable users quickly and efficiently to draw connections between records. Some of these, such as the ability to click on an author and view all records associated with that name, are fast becoming commonplace in archival catalogues. Others, however, are more innovative, providing important case studies in the value of intraoperability within digital resources of this kind. Notable amongst these is our use of attributes to describe the materiality of a manuscript. Matters pertaining to the materiality of the document are outlined in the fields contained within the ‘Physical description’ section of the catalogue, providing a detailed overview of these material features and their place within the archival item. Accompanying these descriptions, however, is a series of keywords, all of which, as discussed above, are clickable attributes with their own corresponding landing page. These attributes cover such features as tears, stains, ink blots and postage stamps, and equivalent keywords have been used for matters pertaining to additions to the manuscript such as marginalia, doodles, pen trials and corrections. Through these linked attributes, users can identify records within the collection that share common qualities. They could, for instance, click on a keyword for ‘ink blot’ and see all the manuscripts within the corpus that also feature this characteristic (Figure 2). They may also filter search results by these descriptors, searching for a term such as ‘Candide’ and applying a keyword filter of ‘Corrections’ to show any manuscript within the corpus that relates to Candide and has been edited. The presence of phenomena such as marginalia, as we have argued elsewhere, is a facet sorely lacking in numerous manuscript catalogues, and the ability to search and filter results according to these facets is unprecedented, making CMV a forerunner in this type of intraoperability and search filtering (Screti 2024).

Figure 2

An example of linked attributes as shown in the beta version of CMV.

In terms of intraoperability, CMV also communicates with a second, sister resource: the Voltaire Library Database. This catalogue of the books known to have been owned or used by Voltaire, also developed by the Voltaire Foundation, uses the same digital framework as CMV and shares a data repository, allowing the two resources to seamlessly interact with one another in meaningful ways. The items in Voltaire’s library known as his pots-pourris, for instance, which are bound collections of print and manuscript material, cross the divide between the two databases. Users will thus be able to navigate between the two resources in such instances, with this linking of the two databases enabling users to consider Voltaire’s literary culture in a broader sense, seeing neither his books nor his manuscripts in isolation but rather as two halves of one intricately interlinked whole, placing the items within a wider context.

At a later stage, the catalogue will also link to the digitised version of the Œuvres complètes de Voltaire (OCV), also currently in development at the Voltaire Foundation, enabling users to access digital scholarly editions of the manuscripts they are exploring within CMV. If, for instance, users are viewing a CMV record for a manuscript that constitutes a draft of a chapter of Voltaire’s Siècle de Louis XIV, they will be linked directly to that passage within the Voltaire Foundation’s digitised version of the OCV edition. Thus, whilst remaining as three distinct resources, the intraoperability between this suite of tools will provide a more comprehensive approach to scholarly research.

The digital form also allows CMV to link out to other relevant resources beyond those developed at the Voltaire Foundation, enhancing the user experience in ways that are impossible to replicate in print. Within a manuscript record, for instance, users are able to navigate out to the catalogue record within its host institution’s website. This will enable users who wish to view any given manuscript to visit the institution’s webpage and order up the item directly. They will also, where available, be provided with a direct link to a digitised reproduction of the manuscript.

The ability to link data within CMV has manifested in far more complex ways, however. Within the data model, we have constructed a series of subsets of data that we have called ‘Collections’, subsets which have been grouped together for one of two reasons: either they represent the component parts of a single physical bound collection of several manuscripts that have been catalogued individually, or they represent iterations of a single text. These have been termed ‘Physical collections’ and ‘Text collections’ respectively. For instance, there are many fascicules that consist of hundreds of letters from Voltaire, such as the bound collections of letters housed at the BnF. Whilst these compendia are often catalogued by libraries as a single unit, it is much more useful to researchers to know precisely which letters are in that bound collection, and what these letters consist of both intellectually and materially. In CMV, therefore, each letter is catalogued individually. However, the physical bindings of these collections are often important elements that shed light on their provenance, and thus it would be remiss to divide up the contents without recognising that the manuscripts are just one part of a wider whole. A ‘Physical collection’ allows us, therefore, to digitally reconstruct the larger volume, whilst still recording each item individually, and, as noted above, these individual items within a collection can be ordered to represent their physical place within it.

The same principles have been applied to conceptual collections, dubbed here ‘Text collections’. Whilst these collections have no physical form, they represent a subset of the CMV data that all concern a particular text, whether that be a literary work or an item from Voltaire’s copious correspondence. Although these types of collections are not physically grouped together, they are linked conceptually by the text that they embody or relate to and thus in gathering them within a linked subset of data, we enable users to easily trace manuscripts that bear witness to the writing, editing and circulation of a given work. The collection for La Pucelle d’Orléans, for instance, collates entries ranging from contemporary copies and translations of the text to letters from Voltaire to his publisher, Georg Conrad Walther, discussing the publication of the work. This functionality thereby offers users a valuable research tool, allowing them to consider multiple aspects of a work from its inception and journey to publication, to its subsequent circulation and reception. The creation of text collections within CMV, we hope, will be of particular value to scholars interested in the genetic dossier as it will allow them to see in one place all extant iterations of a particular work.

The functionality of these collections was one of the greatest challenges in the architecture of the database. The list of manuscripts under each collection is sorted arbitrarily within the WordPress system, meaning that the position of manuscripts within the list is not based on easily evaluable data such as dates, or other standard criteria such as alphabetical order. However, especially in the case of physical collections, we felt it important to display the list of manuscripts in a manner that followed their ordering within their bound format. As WordPress does not natively support this complexity, and as existing plugins available on the marketplace fell short, bespoke plugins were developed to address these specific requirements. Collections, then, represent a significant achievement of the catalogue, and a forward-thinking approach to how researchers may wish to use manuscript materials. A manuscript never exists in isolation and the use of physical and conceptual collections allows us to nuance the complex textual and historical relationships a given manuscript has developed between its creation and its current state.

Conclusion

CMV greatly exceeds, in both scope and granularity, all previous attempts to catalogue Voltaire’s manuscripts. Our focus on intra- and interoperability, enhanced by the depth and breadth of our detailed and granular entries have enabled us to produce a digital resource that allows scholars to engage with manuscripts relating to Voltaire in an accessible and engaging way, supporting both complex searches and dynamic browsing.

In developing a series of new approaches and tools for cataloguing, CMV sets a precedent for what can be done with digital resources of this kind. Voltaire has offered an exceptionally rich case study for building such a catalogue, as his manuscript legacy contains a breadth of diversities and idiosyncrasies which have allowed us to test our model against numerous different cases and make enhancements to ensure accessibility and usefulness. In navigating such diversity, we are confident that the catalogue model we have designed is both robust and widely applicable, ensuring not only that CMV is an accessible and usable resource but also that others can adopt our methods for their own cataloguing projects. With this in mind, our model will be licensable, and we would warmly encourage other cataloguing projects to use CMV’s model.

Whilst the catalogue has been developed with current and predicted research trends in mind, we acknowledge that research is an ever-changing landscape from which new and unprecedented needs will arise. With its robust yet malleable architecture, CMV is well equipped to respond to these needs as they become evident, ensuring that it can continue to be a relevant and useful tool for scholars of Voltaire and the Enlightenment for many years to come. As the project progresses, feedback from users will play a crucial role in ensuring that we continue to offer the best possible experience and make accessible the wonderfully rich corpus of manuscripts relating to Voltaire. We plan, for instance, to allow users to submit feedback regarding specific records directly to us from any given record page should they have additional insights to contribute to the record or should they identify any errors in need of amendment. Broader feedback on the catalogue as a whole is also welcomed. Some feedback has already been gathered: prior to construction, a questionnaire was sent to potential users to identify their expectations of the resource and we incorporated the results of this feedback into the model during the design phase. This type of input will, however, have continued importance for CMV, and we would encourage users to contact us with their comments or suggestions so that we can continue to enhance the project and support access to manuscripts relating to Voltaire.