Posted by e.m.vine@exeter.ac.uk
2 December 2025Emily Vine
Having recently reached a very significant project milestone, we thought this would be the perfect moment to update you on the current status of our research.
Zooniverse Volunteer Phase Complete!
Our big news is that October 2025 marked the end of the Zooniverse phase of our project, which had been running since October 2024. Over the course of the past year, our Zooniverse volunteers have been checking and correcting lines from the 25,000 will transcriptions that had been generated by the Handwritten Text Recognition (HTR) software. On the Zooniverse site, volunteers were asked to compare an automatically generated line transcription with the corresponding image of the manuscript, or to transcribe a line from scratch. Each line would be shown to three different volunteers, to allow us to manually check any disagreements or disparities.
This was very much an iterative process – as the Zooniverse volunteers completed corrections, our technical experts, Mark Bell and Harry Smith, were able to feed this training data back into our HTR models to improve them. We were then able to target additional workflows according to where improvements were needed, such as for some of the difficult sixteenth-century handwriting. We also added workflows which focused on the ‘marginalia’ of the wills – the notes in the margins of the pages which often included the name of the testator, often in a latinised form. Names can be notoriously difficult to decipher, because they often lack the contextual clues that allow us to read and make sense of other words. Having these targeted workflows allowed us to more effectively focus our volunteers’ efforts.
Thanks to our volunteers’ work, we have now reached the stage where our HTR models are accurate enough to generate our final sets of transcriptions. This milestone was met when our Zooniverse volunteers completed checking and correcting all the images that had been uploaded to the site. In total, they checked 436,638 lines, making corrections where necessary, and transcribed 77,761 lines from scratch.
In total, 4,969 users contributed to the site over the course of the year – countless numbers of whom volunteered many hours of their time. THANK YOU, on behalf of the whole research team, to everyone who contributed – this project would have been entirely unworkable without your efforts. Your contributions will enable these invaluable documents to be much more widely accessed and used. We intend to make these transcriptions freely available in the near future, and hope they will be an important resource for a range of different forms of historical and genealogical research for many years to come.
Project Team Workshop and Publication Plans
In a happy coincidence, the final Zooniverse line was checked and completed on the same day as our recent project team workshop, which we held at the University of Exeter at the end of October. We met to discuss our plans for the transcriptions and the database, as well as future publications. Harry and Mark are planning an article about the methodology used to transcribe the wills, focusing on different ways of measuring the accuracy of HTR. Jane Whittle, Harry, and Emily are planning an article that will focus on some of the overarching patterns that emerge when we look at all 25,000 wills, and will explore some of the changes that occurred between 1540 and 1790. Laura Sangha will be writing about the ritual, memory and the material culture of death and commemoration.
We are also planning an edited volume of essays, comprising chapters written by members of the project team as well as invited contributors who will be given access to our database. It will showcase how wills and the project database can be a useful resource for a range of different types of historical research and modes of enquiry. Our contributors will focus on a diverse range of topics, including social status and occupational identity, family relationships, and global trade, as well as in-depth studies of different objects or object types. We used this meeting to workshop ideas, make suggestions about each other’s approaches, and ensure there wasn’t significant overlap between our planned publications.
The project database
Our Research Fellow Harry has been working hard to build our project database, which will allow us to ask questions of the wills, including the prevalence of certain types of objects, and how this changes over time. It will also allow us to group certain types of testators together – so we can look specifically at the objects mentioned in the wills of e.g. women, or widows, or people who lived in Devon, or people who worked as hatmakers.
In the workshop we talked about some of these different categories. We discussed how we might categorise or group objects together: by function (e.g. clothing, tableware), or by material (e.g. silk, mahogany)? We are looking at ways of including both frameworks of classification. We also discussed the different ways of ranking the social status of the testators, and how to determine whether they lived in rural or urban areas. These are all ongoing discussions, but we’re excited to see how the database progresses, and we’ll continue to ‘show our working’ in terms of how decisions are made.
What’s next?
We hope to share preliminary transcriptions in the near future, and, in time, to make other research data available too. We’ll also update the project website and blog with details of any publications arising from the project. You can keep up to date by signing up to our project mailing list or following us on Bluesky. We also update this blog every few weeks, and continue to share full transcriptions, and our thoughts about interesting wills, as part of our regular ‘Will of the Month’ blog post series.