Priests’ libraries – Part 2: making our data operable

by Jaap Geraerts and Teresa Wendel

In a previous blog post that appeared almost three years ago (tempus fugit!), I introduced a small (pilot) project on priests’ libraries from the early modern Dutch Republic. Since then, a lot has happened – in general, but also in relation to this project. Apart from the publication of a chapter in an edited volume on this topic, two research assistants, Sarah Büttner and Teresa Wendel, have patiently and conscientiously identified (most of) the books mentioned in the inventories that form the basis of this project and linked them, whenever possible, to records in online catalogues. Moreover, the latter has created a data model that we developed further in an iterative manner. Ultimately, this data model makes it possible to discern trends within individual libraries and across them, thus furthering our knowledge of the intellectual and pastoral formation of priests serving in the Holland Mission (the Catholic mission in the Dutch Republic), among other topics. This blog post will offer a glimpse of the process of developing a data model and making our data operable in a graph database.

From data to model to graph database

As frequently discussed in articles on the digital humanities (DH), the digital revolution has created new opportunities for exploring and analysing data in the field of historical research. With that in mind, and our processed and prepared data in hand, the focus of the priests’ libraries project shifted from data gathering to establishing a consistent methodology for data modelling and implementing a graph database to facilitate our exploration of the data.

Our journey in this project continued by working with the existing data, which was stored in an extensive Excel file detailing the contents of the various priests’ libraries. This file provides insights into the inventory of the various libraries by listing information such as book titles, authors, dates, languages, and locations, along with a standardized title and the corresponding (WorldCat-)link. To further potential insights, we considered the possibilities of interconnecting the tables, and thus the libraries. In order to do so, we created a lookup table that links different copies of the same book held in different libraries under a platonic identifier, basically capturing the distinction between a (book’s) title and the various copies thereof. Each individual copy of a book, irrespective of whether it’s a translation or different edition, is linked to this platonic identifier.  The population of this lookup table had to be done manually and represented a fair amount of work, further underscoring the labour-intensive nature of the creation of the data that underpins many a DH project. Still, it is worth the time and effort, for in the end this lookup table serves as a reference point for garnering and accessing more in-depth information about our data and opens up new pathways for discovery.

With the process of organizing and standardizing our data being established, our next step comprised the creation of a data model tailored to the specific characteristics of the data of the priests’ libraries. We began by reviewing the information in our tables and identifying any additional information that may be required or added in the future. Once we had identified this information – which ranged from the genre of each book to its format and size, to its owners, donors, translators and publishers – we assessed whether this information would work best as a node, relationship, attribute or label for the purpose of our database, resulting in a preliminary model. After having reviewed the data model, we made several minor alterations to it. The ensuing model connects the platonic books from our created lookup table to their different copies, which were then connected to their author(s), language, publication date, size, etc. In addition, individual copies of a title were linked to the library they formed part of. The following step was to import the data into a graph database in order to assess the functionality of the data model and pinpoint areas that may need further refinement.

The decision to use a graph database was made to focus more on the interconnectedness and analysis of our data. Graph databases, with their emphasis on relationships and traversability, were well suited for exploring the priests’ libraries. Moreover, graph databases are flexible and allow for the easy extension of the data model that is used, making them perfectly suitable for an evolving project like this one. Using Neo4j, we translated our data model into a graph structure, with nodes representing entities and edges capturing their relationships. The result was a database with around 3,700 nodes and 8,500 edges. We now arrived at the point where we can start querying the data and continue the refinement of the data model, enabling further routes through which the data can be traversed and thus opening up other avenues of research.

A simple query reveals the possibilities of the data model as it currently stands. A logical starting point forms Cornelius Jansenius’ Augustinus, the posthumously published book that set off the so-called “Jansenist controversies” that caused such deep rifts in the early modern Catholic Church. In the Dutch Republic, these quarrels led to the emergence of the Church of Utrecht (nowadays the Old Catholic Church) that competed with the Roman Catholic Church in this officially Reformed Protestant country. As one would expect, several clergymen who were part of the former Church owned a copy of Jansenius’ Augustinus, but so did the Jesuits in the city of Haarlem. In itself this finding is not shocking – some priests’ libraries also included books written by Protestant Reformers, if only because one had to know one’s enemy – but it shows the potential of being able to search across libraries. What mark did the “Jansenist controversies” leave on the libraries owned by (Dutch) priests of various colours and stripes?

Moving forward

When considering our methodology, it is evident that our focus extends beyond executing queries or generating insights. Yet so far our work has centred on the data itself. We curate, organise, and connect data, recognising that the accuracy, depth, and richness of the data stored in the graph database foster the extent to which this project can stimulate new and exciting research. In particular the comparative angle that is made possible by this data model represents a significant advantage in comparison to having to compare printed or manuscript library inventories in an analogue context. Hopefully, we can acquire funds to further extend this project in the future, for example through the inclusion of more priests’ libraries and, possibly, the libraries of Catholic laypeople, thereby enabling a more systematic analysis of the books owned by Dutch Catholics as well as wider intellectual culture of early modern Dutch Catholicism.

Cite this article as:  Jaap Geraerts and Teresa Wendel, “Priests’ libraries – Part 2: making our data operable” in Digital Humanities Lab, 22/03/2024, https://dhlab.hypotheses.org/?p=4597.

Featured image: Books owned by the priest Joannes Blo(c)k.

Please use the citation reference above. The reference below is generated automatically and does not take multi-authorship into account.

---

---