Deciphering the Text: the Process of Transcribing SP9/99

Project leads Paul Pattison and Esther van Raamsdonk give an update on how their project on transcribing and translating of the Seventeenth Century survey by a Dutch Engineer of 22 castles and fortifications.

Now we are reaching the end of our research project – Transcribing and Translating SP9/99: A Seventeenth-Century Dutch Survey of 22 English Castles and Fortifications – we wanted to share a little more about the process of coming to grips with this engrossing manuscript.

Subscribe to our quarterly newsletter

A brief bit of context: the National Archives holds an anonymous manuscript, currently undated but certainly early seventeenth century, which contains a survey of at least 22 English castles and fortifications. This survey was carried out by a Dutch engineer, who clearly spent considerable time in England, as he has adopted several English words (albeit with idiosyncratic spellings). Because of the linguistic and material challenges – the unorthodox Dutch, the difficult handwriting, and the bad condition of the paper – it has never been transcribed and translated. The value of understanding the manuscript, however, is clear. The survey outlines the condition of the castles and fortifications at the time; it provides early modern names of buildings and their elements, some of which are now lost. It also provides suggestions for improvement, some of which we know have later been realised. We can now relate these improvements to the suggestions of the survey. More will be said in due course about some of our findings concerning what we can learn about the history and development of the 6 particular castles that we have now transcribed and translated – Sandown, Deal, Walmer, Dover, Sandgate and Camber – but here we wanted to elaborate a little on the fun and challenges of transcription.  

The engineer was from the Low Countries. It is difficult to pinpoint exactly where, but it was likely the southern end of the Netherlands, as the language does not follow the more standardised version of Dutch that by that time flourished in the North. Beyond the language, there is no punctuation in the document, not even a full stop. Almost all text is in phrases; there are no complete sentences; frequently verbs are missing. It reads like a list or a summing up of the engineer’s thoughts as he examined the sites. However, the content is also careful and precise, noting measurements, directions, costs and uses, all accompanied with skilful drawings and plans.

Subscribe to our quarterly newsletter

We will give you here two examples from Camber Castle. We have included the engineer’s full sketch plan here but will zoom in on the central keep. Having spent several months with this engineer and his handwriting, it can feel like we know the man quite well. He was often in a rush. This is visible in the way he sometimes repeats the same word in a row, and the density of abbreviations used. In the early modern period, abbreviations were common, and were themselves often standardised. For example, yt for that, or Sr for sir. These could be marked in several ways, but most often in either superscript or with a line above the word. Our engineer liked to break with tradition and places a characteristic C above a word. In my years working as palaeographer, I have never come across this, so we can reasonably conclude that this engineer was not formally trained. That is to say, he did not go to university; for most of the early modern texts that we work with, the authors had received a formal education and conform to the ‘rules’ of writing, but our engineer had worked out his own system of noting deviations. In a similar break with tradition, he uses the C symbol in three ways: to mark an abbreviation, to flag where he has made a mistake, or to indicate an English word and that he does not know how to spell it. Which one of these is the case for individual words is up to us to find out.

Image of Camber Castle from manuscript SP9/ 99. Copyright Paul Pattison

In the image below, there is a good example of the difficulty of these abbreviations or deviations. In the keep of Camber castle, there is written ‘ende boven, …… 3r 1v’. We know there is something strange going on with the word ‘boven’ because of the symbol C. To confuse matters further, this handwriting uses the same letter for v and n, and its e is often written as an o. We therefore assumed that this word would be an abbreviation of ‘benen’ (sometimes written as benéen), meaning ‘beneden’, which is Dutch for beneath or underneath. However, after working on several castles, it did not make sense that he would be talking about a roof beneath a room, and we therefore had to revisit all instances of benen and realised in some cases it had to be ‘boven’, meaning above. In this case the C symbol merely signifies the confusion between n and v, that the engineer himself clearly also suffered from.

Camber Castle image close up from SP9 / 99. Copyright Paul Pattison, National Archive Kew

To give an example of the engineer’s haste, or possibly enthusiasm, we can turn now to the marks just above the word ‘boven’ in the previous image. As is clearly visible from the full photo of Camber castle, this was a gifted draftsman. Camber Castle is certainly not the most technically difficult of the castles in the manuscript, but, as ever, his attention to detail and scale is impressive. The care he took in the drawing itself is not always present in his annotations, presumably because these were in draft form. As a result, the distinction between what is part of the drawing and what are notes that were added later is blurred. In the below image on the left, there is a strange mark which can be transcribed as ‘hffo’. This is not even close to an early modern Dutch word. The h also misses its characteristic full loop in the bottom curve. We thought it might be an abbreviation, although the C-symbol is not present, or some sort of mark highlighting an element of the building. It was only after we went over the full transcription several times again, that we realised he had misspelled ‘hoff’, like the word on the right, meaning ‘courtyard’ in Dutch. In his rush he had jumbled the order of the letters and not fully closed his h.

Camber Castle image close up 2 from SP9 / 99 at National Archives, Kew. Copyright Paul Pattison

These moments of breakthrough are rather exhilarating. Most of the transcriptions and translations that we have made still have some outstanding ‘curiosities’ to be solved. However, as a result of these idiosyncrasies or oddities, we have been drawn very close to the material and grappled with all aspects of it: the material history, the background of the engineer, the process of surveying these castles and fortifications, and the long history of repair. We are now in the process of puzzling over the new information the survey has brought to light and how it fits with what we already know about these buildings. We look forward to disclosing further updates here, hopefully in fully legible modern English.

Subscribe to our quarterly newsletter

A Dutch survey of English Castles and Fortification in the seventeenth century

The Castle Studies Trust has awarded a research grant to Dr Esther van Raamsdonk (University of Utrecht, Netherlands) and Paul Pattison (English Heritage) to work on an early 17th-century manuscript from the National Archives, concerning fortifications and castles on the south coast of England.

Subscribe to our quarterly newsletter

The manuscript was discovered over 40 years ago by Charles Trollope, following which, the late John Kenyon published a preliminary assessment (Fort vol 11, 1983, pp 35-56), showing that it related to a survey of several coastal fortifications along the south coast of England, mainly those that originated in Henry VIII’s castle building programme in the 1540s. However, the manuscript is in Dutch, so as Kenyon rightly says: ‘translation of the Dutch is required to assess the document’s full significance and to possibly produce a more exact date than I am able to postulate at this moment’ (p. 37). His article examines some of the plans in the document, but there is no engagement with the Dutch text, which is more than 50% of the document. No transcriptions or translations have therefore been made before.

Sketch of Walmer Castle by the Dutch Engineer, courtesy of the the National Archive

The manuscript seems to be a surveyor’s on-site notebook, comprising fairly accurate and proportioned sketch plans. These beautiful drawings show a highly-skilled engineer at work, and one with a talent for exact survey drawings. These are supported by keyed notes and more detailed explanatory text. It covers most of the main fortifications and castles along the south coast, from Sandown Castle near Deal in Kent to Pendennis Castle at Falmouth in Cornwall.

Walmer Castle today copyright English Heritage

Paul came across the manuscript several years ago and recently, prior to writing a new guidebook for Yarmouth Castle, Isle of Wight, looked for a palaeographer to transcribe and translate the pages for Yarmouth – enter Esther. Paul is a historic fortifications specialist, and his job is to help to make sense of the translation, as the conversion to English is not always obvious: there are many military and technical terms and sometimes the meaning is not immediately apparent. However, the sketched plans are annotated and keyed, which should help in understanding specific parts of each fortification. Esther is an expert palaeographer, having taught early modern English palaeography courses for some years, and scholar of 17th-century Anglo-Dutch relations. The grant is towards her costs of transcribing and translating the text and notes. It is not an easy task. The hand is a difficult one and several pages are faded. The Dutch engineer spent some time in England and the result is a hybrid language, a phonetic (for Dutch readers) spelling of English words. The document also functioned as a shorthand for the author, using non-standard abbreviations, idioms, and little drawings. The exercise is therefore not of a straight-forward transcription but requires quite a bit of puzzling.

Subscribe to our quarterly newsletter

As mentioned, we used the Yarmouth example to see what information we could gather from the manuscript. From this test case we know that the surveyor was commenting on condition and making recommendations for repairs and changes. Because of the detailed map and key, we could pinpoint the function of some rooms that were hitherto unknown. We look forward to see what further secrets the document will reveal about the other buildings.

The funding will enable us to examine a sample of the manuscript covering several sites in the Cinque Ports area of jurisdiction – Sandown, Deal, Walmer, Dover, Sandgate and possibly Camber. All of these are well known to Paul, and it is hoped that significant details about the condition of each castle or fortification will be forthcoming, for comparison with other contemporary accounts in English, to see and understand where this survey fits and whether or not it resulted in any repairs or additions on each site. John Kenyon thought it might date to the 1620s: we would like to try and confirm that, or otherwise, by comparison with earlier and later surveys, and to establish the political context for its commissioning. This will lead to at least one provenance article. The transcriptions (some of which will be published) will also provide a very valuable snapshot of the south coast fortifications 80 years or so after their creation, and form a test for the usefulness of going further and transcribing the whole manuscript. There is an exciting journey ahead.

Subscribe to our quarterly newsletter