Martin Cook has sent us this guest post on his suffolkwills.org project and its use in studying monuments and burial practices

Using wills to study monuments and burial: suffolkwills.org

Let’s say you had the question, “How many people in Clare, Suffolk, wanted a monument for themselves between 1500 and 1550?” You want to search wills for this, but how? Sites like Ancestry.com allow you to search wills by proper names, date or place, but aren’t made to answer general questions about charitable giving, household goods and monuments. Their records don’t contain this information.

In creating suffolkwills.org, I digitised Tudor wills and developed a template (or schema) for describing them, so each will has a detailed record about it. The records contain, for example, the animals mentioned in a will, the clothes mentioned, and where the testator wanted to be buried. We can then search and analyse these records to answer many different questions.

This article describes how the site was built and gives examples of what it has revealed so far about burial and commemoration in sixteenth-century Suffolk.

The scope

Since I was also interested in how wills changed from the pre- to post-Reformation period,  I restricted the time scope to the sixteenth century.

I kept the geographical scope to Suffolk, because many Suffolk wills from Tudor times have already been transcribed, and are in the Suffolk Archives. The transcriber was Peter Northeast, a renowned Suffolk historian who transcribed many primary sources from the county. The archives gave me permission to digitise the transcriptions, and put them online.

The method

1. I took photos of the will transcriptions in the Suffolk Archives. The transcriptions are handwritten on A4 paper.
2. I asked AI to convert these transcriptions into digital text. I then compared the text that it produced to the original handwritten transcription, to make sure the digitisation was accurate.
3. I then developed a schema to record each will. A schema is a template for the information you want to record about something. It defines the fields that can be filled in for each will, which fields are required, and what type of data each field should contain. For example, the will schema includes fields for the first name, surname and gender of the testator. It specifies that the “Date of will” and “Date of proof” fields must contain the date format YYYY-MM-DD. The schema also tells you if the field has a restricted list of options (a Controlled Vocabulary), so the “Marital status” field, for example, can only be “Married”, “Single”, “Widowed” or “Unknown”.
4. I asked AI to put the digitised text into the schema. In other words, I asked it to pick the elements of the will text out and put them in the appropriate fields, so that the first name would go in the first name field, the last name in the last name field etc. Again, I checked each record manually to make sure it was accurate. Since the schema is so detailed, this step would have been too painstaking without AI.
5. I put the records of the wills (and the will texts) on a website.
6. I then wrote some scripts to answer questions about the wills. These scripts generate the tables and graphs in the analysis section of the website (they don’t use AI). They answer questions like, “How many people wanted to be buried inside the church rather than the churchyard?”, “What were the most common household goods mentioned?” and “How did soul commendations change across the 16th century?”

This was not a linear process. It involved repeatedly looping back and modifying the schema to fit the data, and updating the records already processed to fit the updated schema. For example, I was not able to find a schema to describe a Tudor will, so I created one. This meant guessing the fields needed for an initial schema, recording wills using this schema, and discovering fields I needed to have, but didn’t. Then I added the new fields to the schema and updated all the wills I’d already processed so they had these new fields.

It was not only recording wills that led to altering the schema, but also their analysis. For example, I wanted to ask, “Which relatives were household items most commonly left to?” I had a field for “Household items”, which had a “Recipient” sub-field, but not a “Relationship” sub-field. I may have recorded that Marjorie Wade received a silver spoon, but I hadn’t recorded who Marjorie Wade was in relation to the testator. So I added a relationship field, which allowed me to answer that query.

Why not use Language Learning Models (LLMs)?

LLMs are a type of AI that specialise in human language. You could say, “Wills are quite formulaic, why not ask an LLM to analyse the wills directly? Just give it digitised wills and ask it questions like ‘How many people wanted monuments in a church?’. Then you don’t have to create a schema.” The answer is:

• Auditability and repeatability: with a schema it is clear how the result has been calculated, and you can trace every result back to the wills that produced it. If anyone wanted to check the results, they could run the same analysis using the same schema. How an LLM arrives at a particular answer is often less clear.
• Interpretation: some of the phrases in the wills are ambiguous and need some knowledge of context. For example, the phrase “to be buried in the church” did not always seem to mean “in the building of the church”, since people used it who couldn’t have expected to be buried inside the church. It appeared to sometimes mean “in the holy grounds of the church”.
• Consistency: a schema forces judgement calls about ambivalent items. For example, are rosary beads jewellery, accessories or even household items? With a schema you can choose one category to assign them to, give the category an explicit definition and use it consistently. I have found AI can vary in its interpretation of items, even after you have told it what category to use.
• Reusability: once the records for some wills have been made, they can be re-analysed in different ways without having to re-read the wills again. An LLM may need to re-read and re-classify the wills every time a new question is asked.
• You don’t know when something is wrong: checking the will texts, and checking that AI has filled in the schema for each will properly, gives you an intuition of what is typical in the wills. If a result has been miscalculated, you have a better sense that it is wrong.

Having said all this, the more manual approach is only practical for a smaller data set. It would be impractical to use it for large digitisation projects (such as the Material Culture of Wills Project).  Also, AI is useful for spotting patterns across all the wills, like repeated names of people or properties, so I plan to use it more for this, though perhaps combined with traditional programs.

Monuments and burials

So far the website has only a small sample of wills (around 220), but what have the wills suggested so far about monuments and burial? Let’s ask a couple of example questions: where did people want to be buried? Who supervised the design of the tomb?

Where people wanted to be buried and why

Just over 50% of testators said they wanted to be buried in the churchyard of the parish church. About 20% wanted to be buried within the church, or within a monastery church, 12% left it up to their executors or “where god pleases”, and 13% did not give instructions about burial location at all. It could be just “To be buried in Christian burial”, for example.

People who wished to be buried in the churchyard usually give no specific location. Sometimes the testator asks to be by a loved one, like Martha Erensby of Cockfield (1566) who requested, “To be buried in the churchyard there, by my husband.”

As the century progresses and Protestant beliefs become more apparent, then there is a suggestion that the grave is just a place to await Judgement Day, so the location is less important. Here is William Gorstlen of Aldeburgh in 1567, for example: “As for my body, I will it to be buried, not making mention of place, believing that in the general resurrection of the dead that I shall rise an incorruptible, an immortal & glorified body & shall see God my saviour, and so both soul & body being united together shall be placed in the kingdom of heaven. “

As for church burials, presumably in pre-Reformation times the testators were hoping that their grave would prompt prayers, though this is usually not made explicit. One exception is James Goldwelle, Bishop of Norwich, who appears to have built his grand tomb before he died (still in Norwich Cathedral), and who wrote in his 1497 will, “I wish my exors. or their deputies to distribute each Sunday for 3 whole years next & immediately following my decease, to 20 poor men 6s 8d & that distribution to be made when the bell rings at high mass & afterwards they to go to my tomb & there pray for my soul.” So he conceived his  tomb as a focal point for intercessionary prayers.

Around half of the secular people who wished to be buried inside the church wanted to be next to a loved one. Usually this was their partner, but in one case it was the testator’s uncle and in another his sister.

In 1502 John Hadley of Clare, for example, wants his “vile body to be buried within the parish church of Clare next by the sepulcher of my wife”. A century later (in 1621) Lady Ann Drury wants her “bodie to be laid in the tombe at Hasted [Hawstead] by my deere husband and dawghter Dorothie.”

The partner may not be dead, however, and here the pew where the partner sits provides the burial location. Edwarde Gelget of Ipswich in 1521, for instance, wishes “To be buried in the church of St. Mary Keye before the pew of my wife”. Sometimes the pew itself seems to be the focus, as in John Grene of Beccles (1531): “To be buried in Beccles church, near unto the place whereas I do use to sit.” His wife was still alive, and presumably she sat there with him, so it’s curious he doesn’t mention her.

The design of the tomb

Only two testators mention that they have already prepared their resting place: James Goldwelle, Bishop of Norwich, who we have already mentioned, and John Persay, a draper of Bury, who in 1509 wanted his “body to be buryed in Seynt Mary chyrch vndr ye grave stone off marbyll ther by me layd jn ye sowth yle off ye sam chyrch.”

John Coote of Bury in 1502 provides a rare instance of someone who wanted to be buried in the churchyard, but went into details about his grave. He asked that “aft my decesse to haue ij crossys goodly of tymbyr, one at my hede the tother at my feete, wyth my armys thereupon, and also a wrytyng to pay the peopyll of ther charite to pray for my soule.” He is also an example of someone who specified the burial location within the churchyard. Again, it was near loved ones – next to his children and a priest.

Future questions regarding monuments and burial

Additional interesting questions for the future would be  – did people get their wish, especially if they asked for a location within a particular church? We know that Dame Elizabeth Barnardiston in 1526 asked to be buried at Walsingham Priory, but ended up in Kedington church. From the churchwarden records of Mildenhall, we also have the case of the vicar there John Wilkinson, who died in 1541. He owed some lead (or its worth in money), and when he died the stone that was prepared for his grave was sold to pay off his debt. This question will involve research beyond the wills, of course, but the wills can be a starting point.

As more wills are added to the site, more subtle questions can be asked about monuments and burial – for example, around the funerary rituals. Here is the priest Sir William Paynter from Bardwell, who in 1559 knew things had changed under Elizabeth I, but holds out hope for a traditional burial when he asked, “yf the lawes of the realme do serve and the procedyngs of the heyghe powers wyll suffer by the order of the lawe, to haue the obseruatyons and ryghtes of the catholyke churche.”

Conclusion

The Suffolk Wills website was started in December 2025, and so is still in early days. Hopefully as it grows, the analysis of the wills will give a clearer picture into aspects of sixteenth century life in Suffolk, such as burial and memorialisation. People are quite welcome, also, to suggest analyses that they would like to see on the site.

Going forward, would it be possible to create a simple schema for church monuments, to record monuments in a consistent way and analyse them across regions and time? I hope the Suffolk Wills project can give ideas about how that might be done.

Tips on using the site to find potential monuments

• The Burial page in the Analysis section lists people who wanted to be buried within the church.
• Any testator who mentions wanting to be buried within the church also has the keyword “monument” added to their will record. You can search for “monument” from the home page search box to find these wills. Similarly, the word “marble” is added as a keyword to wills that mention a marble stone for a monument. You can search for “marble” to retrieve these wills. All wills with the keyword “marble” will have the keyword “monument” as well, since the testator wished to be buried in the church. Note that keywords have modern spellings (no need to write “marbull”, “marbyll” etc).
• If you want to restrict your search to a place or time period, go to the Wills section and open up the “Advanced search” box. Put your keyword (e.g. “monument”) in the “Search terms” box, then select your “From year”, “To year” and “Location”. The “Location” field only contains the places where the testators in the will collection lived. More places will appear as more wills are added.
• None of this guarantees, of course, that there is a monument in the church mentioned, or that the testator was even buried there!