Tracking the life of the book: an interview with Cristina Dondi Free

The methodological shift, applied to incunabula already described at the bibliographical level in the Incunabula Short Title Catalogue (ISTC) database ^[2], consisted of three components:

1. Detailed copy-specific cataloguing: Combining internal evidence (ownership inscriptions, manuscript annotations, decoration, binding, etc.) and external evidence (historical library catalogues; sale and auction catalogues; sequestration lists, etc.).

2. Creation of the international database MEI^[3]: Designed to keep the data together, allowing researchers to reconstruct dispersed collections and track the movement of books from the 15th to the 21st century.

3. Promotion of collaboration and shared knowledge.

The continuous development of CERL’s digital resources is our top priority and the focus of our financial investment. The strategic priorities in digital scholarship are:

• Data integration: Promoting and ensuring the integration of all bibliographical and copy-specific data for European and American printed heritage up to circa 1850 into a specialised network.

• Enhancing access: Supporting access to this data through the CERL Thesaurus (which includes variant spellings of imprint places, personal and corporate names) to enable high-quality and comprehensive scholarly research.

• Preserving bibliographic heritage: Maintaining the bibliographic heritage of Europe, including incorporating large international resources such as the ISTC, STCN and ESTC, with the goal of improving access and interoperability.

• Extending the MEI model: Extending the model created for copy-specific data collection and analysis with MEI to other research projects (e.g. HPB-Provenance, Patrimonit, Matmed) to ensure long-term sustainability and improve interoperability within the digital environment ^[4].

To address the need for global accessibility and interoperability, CERL is progressively making the data in its digital resources available as linked open data, beginning with the Thesaurus.

Freely available digitisations have been immensely useful, advancing incunabula research in multiple scholarly ways. Furthermore, I have made extensive use of digitisations for the creation of many videos displayed in the exhibition “Printing R-Evolution 1450–1500” ^[5]. Since then, all the digital material has been made available online and is regularly used for teaching and presentations, with libraries also running them in their exhibition spaces (for example Monastery of Santa Scolastica in Subiaco).

In Oxford, during the 15cBOOKTRADE Project ^[6], we experimented with image matching of illustrations in incunabula to help find when woodcuts were reused and copied by different printers in different places and time. This year we plan to apply image recognition to the images of provenance contained in CERL’s Provenance Digital Archive (PDA) ^[7]. But to achieve high quality advancement in research every implementation of AI requires a vast amount of (new) data collection and data organization. At the core of any exciting possibility stands good quality cataloguing, and I am very concerned by the pockets of important material not yet catalogued in our libraries.

At present I teach, in English, History of the Book (specifically the introduction of printing in Europe and around the world, in Latin and non-Latin alphabets) to a very diverse and large group of first year undergraduates within the Global Humanities BA course; they come from all over the world. I regularly use the videos created for the Venice exhibition and for other projects I have been involved in, as well as the digital resources we created (MEI, PDA, 1516, Text-inc) to explain the historical objectives which we are working towards, with a strong emphasis on methodology. But I also take them to libraries, to the Tipoteca Italiana and to Fabriano to experiment with the materiality of printing.

Students should have a direct experience of the object of their studies, that is books and libraries; contributing to ongoing research projects or large cataloguing campaigns and internships are very important; these experiences should reinforce their knowledge of the existing main digital resources as provided by teaching; and they should speak English, to benefit from international mobility.

Everything CERL hosts and curates is made freely available; for the political aspects we conform to European legislation.

For us at CERL, another concern is version control. Scholars annotate and enrich the data they download from libraries/large repositories. When, in the not too distant future, scholars begin to make use of the Cultural Heritage Cloud, and themselves upload data there, eventually it will become very hard to understand which version of the data is available, who created it and precisely what the intellectual status of that data is.

One measure of success of our digital resources is the willingness of libraries and scholars to contribute to them – an investment of time, expertise and money. Other measures include the frequency of citation, invitations to speak at both scholarly events and those for a more general public, and scholarly publications and articles written for a wider audience. Then, of course, there is success in acquiring funding for research and special projects (for example, Dante1481 ^[8]; the digitisation of incunabula in Italian monastic libraries ^[9]; the digital reconstruction of the library of San Giorgio Maggiore of Venice, which is ongoing).

In terms of strategy, high-quality cataloguing – not in a silo but as part of a connected network – is always the core component of our research projects. However, in research projects, high-quality cataloguing should always be accompanied by a scholarly output and effective outreach, such as videos and exhibitions. Communication at multiple levels is essential.

When I started MEI in 2009, I had the good sense of asking the Data Conversion Group of Göttingen State and University Library to design it with me. They have designed, managed and further developed most of CERL’s digital resources, in collaboration with the librarians in CERL. The collaboration with, and support of, CERL libraries is part of the success of MEI and other CERL digital resources. I worked closely with a number of engineers specialised in data visualisation and image recognition (Min Chen, Andrew Zissermann and the Visual Geometry Group in Oxford; Alfie Abdul Rahman and the team of computer scientists at King’s College London). Clear and close communication with the scientists is vital. It is also very important, in my experience, to have clear historical objectives to begin with.

CERL adheres to internationally defined data structures and cataloguing practices, applying them to historical material. It ensures consistency and interlinking by promoting the use of standards and persistent identifiers, thereby gathering rich, specialised information and avoiding, as much as possible, the duplication of effort and loss of valuable data. Furthermore, clearly defining the data sets is becoming increasingly important so that researchers are fully aware of their contents and limitations.

All data added to CERL datasets are derived from trusted sources (rigorous cataloguing undertaken at CERL member libraries, to agreed standards). The data are carefully analysed to harmonise them from various cataloguing traditions and are ultimately supported by a coherent model and stable database environment that offers tailored indexing, searching and browsing. This illustrates the many steps and processes involved in transforming information and knowledge into usable data.

The most rewarding project was undoubtedly the creation of MEI, following ten years of detailed incunabula cataloguing at the Bodleian Library. Despite the wealth of data, the work to extract information on the European circulation of Venetian books from the six-volume catalogues for an article (Dondi, 2008) proved extremely arduous. This difficulty led to the conclusion that the use of incunabula as historical sources could only be improved by reorganising copy-specific information (such as provenance, binding and decoration). From this, the model of the “life cycle of the book” was born, tracing its movement as a sequence of geolocated provenance blocks.

The digital resources which we have or are creating have to be comprehensive to sustain historical research. This is a long term aim. Therefore, I would like to stress again the all important awareness that not everything has been catalogued and even less has been digitized. And if it has been catalogued, it may not have been catalogued to the level of detail that you require for your research. Therefore research needs high quality cataloguing of the heritage collections, and the availability of images of the books, preferably linked to the records. Young scholars should be aware of the vast quantity of books in our libraries; if they are looking for a topic for their MA or PhD they should go and talk to their local librarian. Provenance research (for which high quality cataloguing is required) is essential for the understanding of our entangled and highly mobile heritage collections; although increasingly wide-spread, there is so much still to do, including the integration of provenance data pertaining to GLAM heritage material.