Using Digital Epigraphy Methods for the Recording of Erased Text on Papyri
Written by Elena Luise Hertel, Basel University, Ph.D. candidate, member of the project Crossing Boundaries: Understanding Complex Scribal Practices in Ancient Egypt
The Museo Egizio in Turin is home to many famous pieces of ancient Egyptian art. Less known is that the collection also contains roughly 9000 fragments and approximately 230 larger ensembles of New Kingdom hieratic papyri from Deir el-Medina, many of them unpublished . In order to gain a better understanding of this rich but highly fragmented material, the project Crossing Boundaries: Understanding Complex Scribal Practices in Ancient Egypt was created as a collaboration between the Universities of Basel and Liège and the Museo Egizio. The goal is not only to study the papyri within an international and interdisciplinary research project  but also to provide access to the material  (see the article of S. Töpfer  and create modern, digital tools for facilitating papyrological work .
P. Turin Cat. 1906 verso primary and secondary text (1)
My Ph.D. project within Crossing Boundaries is focused on the use and reuse of papyrus. It is generally known  that papyrus as writing material was occasionally reused, for example by erasing a previously written text and re-inscribing the papyrus (and thus creating a palimpsest), by using the empty back side of a sheet, or even by making use of any empty space such as a margin. However, many questions on the exact dynamics of this practice remain open: Were there specific rules or common practices of papyrus reuse? How exactly was ink removed? What kind of texts were erased in favor of a new text? And how many times could a papyrus sheet be reused?
P. Turin Cat. 1906 verso primary and secondary text (2)
While it might not be possible to answer all these questions with certainty, one promising approach is to take a (physically) closer look at palimpsests and erased texts. The material from Deir el-Medina in Turin includes several papyri showing different traces of reuse and thus makes for an ideal place to start the study. During a two weeks long research stay at the Museo Egizio, I studied a number of palimpsest papyri. One of the primary goals of this visit was to develop a method for documenting what is left of the largely erased primary inscriptions of the manuscripts, the results of which I am happy to present here.
P. Turin Cat. 1906+C2047+C1939 verso (scan copyright by Museo Egizio)
P. Turin Cat. 1906 verso detail (scan copyright by Museo Egizio)
The papyri in question all belong to the large corpus of manuscripts, partly acquired in 1824 from the antiquities collector Bernardino Drovetti and partly excavated directly in Deir el-Medina by Ernesto Schiaparelli at the beginning of the 20thcentury. Most of the papyri I chose for my study – based on the amount and quality of the erased text – are inscribed with texts containing documentary content such as accounts, lists, and receipts. They are written largely in black ink, but in many texts red ink is used regularly, too. As an example, I selected a large papyrus, P. Turin Cat. 1906+C2047+C1939 containing several accounts dated to the regnal years seven and eight of Ramesses IX . The numerous documentary texts on the manuscript were less of a concern to me than the traces of erased signs faintly visible all over the papyrus. Even though the scans provided by the Museo Egizio already give a solid impression on where a previous inscription might have been, my goal was to create a thorough documentation of all the remaining ink of the erased text in order recover as much information as possible from this first inscription phase.
Problems and Challenges
The documentation of the erased texts on the Turin papyri has several problems connected both to the material itself as well as the way they were treated and kept in modern times. Unfortunately, scientific methods employed in the research of palimpsest manuscripts from later times in history – like medieval codices written on parchment, where a primary text can often be made perfectly visible with the help of UV light sources – are not a viable option in the case of our papyri. One reason for this is that the material properties of text and text carrier make it unlikely that such an undertaking would yield any results. Medieval codices are generally written with iron-gall ink of which some components (namely the gallic acids) penetrate deeply into the parchment and can be made visible with UV light, even after the pigments that are visible to the bare eye (namely the metallic parts etc.) have been removed. In our case, however, both the impermeable structure of the papyrus fibers as well as the lack of a component similar to the gallic acids that would survive the erasure process, yet be responsive to light outside the visible spectrum, makes it unlikely that such techniques would reveal much of a previous text. Additionally, the necessary removal of the glass frames under which the papyri are preserved at all times might be possible only for a few papyri and only under special circumstances, thus making this method unsuitable for a large-scale study.
Further, most of the papyri underwent several stages of restoration and conservation procedures of which some date back to more than a hundred years ago. Especially the excessive use of glue as well as a material referred to as ‘gauze’ (a thin, net-like textile made out of silk and used as backing to keep the fragments in place) often makes it impossible to photograph or scan the papyri without reflections.
Microscopic view of P. Turin Cat. 1975 verso with historic restoration (left) and P. Turin Cat. 2003 verso with gauze and ink remains (right)
The use of a digital microscope (Dino-Lite AM3113T) proved helpful for recording close-ups of specific sections, yet not suitable to document traces over larger areas. Therefore, I opted for drawing as the most promising way of documentation and developed a workflow inspired by the digital documentation practices developed by Epigraphic Survey artist, Krisztián Vértes.
1 – Sketching
When starting to work on the originals at the Museo Egizio, I soon realized that all my previous attempts to draw the erased text based on the scans I had were of very little use. Seeing the papyrus right in front of me gave a very different impression than studying the digital images, so I ended up discarding most of my previous work and starting from scratch.
P. Turin Cat. 1906 verso sketch and comments
Loosely following Krisztián’s method, I began with sketching the traces I saw and commenting on uncertain or noteworthy features. I used an iPad (6th gen) and Apple Pencil (1st gen) and the software Procreate. The small size of the iPad proved useful – between laptop, magnifying glass and lamp, Dino-Lite, and a large papyrus frame there was hardly any space left on the desk in the study room. The iPad and software lacked the processing power to work on the entire images (even at only 400 dpi, the files were quite large due to the size of the papyrus) which I bypassed by cutting the image in smaller sections and using those as basis.
P. Turin Cat. 1906 verso result after inking process - detail (Click photo to discover in higher resolution)
P. Turin Cat. 1906 verso result after inking process (click to enlarge)
Not having had any previous experience with drawing ink remains with Procreate, I finally chose the pre-set Procreate Pencil to be found in the Sketching folder (click to download) since I found it easy to handle due to it being set to imitate a physical pencil. I used the pressure sensitivity and the angle of the pen to create lighter and darker areas in order to indicate where the ink was preserved better or worse, and tried to show parts where the signs were fading away by keeping the texture very light. Closer observation of the ink traces revealed that the edges of the brush strokes were noticeable in some occasions, being slightly darker than at the centre of the stroke. Since this information is important for identifying a sign, I decided to record line edges by drawing harder, darker lines. In a way, this is similar to recording damage in a relief: knowing which part of an image (or sign) is intact and which is disturbed is necessary to be able to reconstruct the whole.
I indicated red ink by using the same pencil and technique in a red colour that appeared similar to the colour of the ink.
Notes on uncertain parts, traces looking considerably different on photograph and in reality, details to check again with the microscope, and other useful comments were recorded on a separate layer.
2 – Inking
A screenshot of my multi-layered digital inking workflow developed in Adobe Photoshop
Even though my drawing on the iPad already illustrated the primary text much better than the scan, I decided to try and improve the results by adding another work step in Adobe Photoshop, similar to the inking process in traditional and digital epigraphy. The aim was to render the drawings clearer and more homogeneous as well as to increase the resolution of the end result.
I defined two brushes (one for line edges, and one for internal parts of a line) which should give the drawing a neater look than the Procreate Pencil, yet still keep the character of the irregular ink traces. After a few trials I noticed that the brush technique still led to a rather patchy looking result and distracted the eye away from the shapes and towards the darker spots where more ink is preserved. To avoid this, I decided to work with a pattern rather than a brush, inspired by the method Krisztián developed for recording the late-Roman murals at Luxor Temple. I used one of Adobe’s pre-defined artist’s brushes which imitated dry paint to create an uneven, yet homogeneous looking pattern which I then applied with the Pattern Stamp Tool. Still, along the lines of the digital inking technique developed for the Luxor fresco, I set up three layers – each with a different level of opacity – and started tracing my penciling using the layers for three different levels of ink preservation (dark, medium and light). For marking visible sign edges, I used a standard soft brush and a new layer. Signs written in red ink I recorded in separate layers, (though still following the same method of patterns and layers as with black ink) and colorized the entire layer groups by using a fill layer with a clipping mask.
Comparison between scan (left), sketched version (middle), and inked version (right) (scan copyright by Museo Egizio)
In sum, I think this method leads to a detailed, yet clean result. In this case, access to the original proved absolutely necessary in order to be able to correctly distinguish between ink remains and other darker spots on papyrus and image. Even though the ‘inking’ process added an additional work step, I consider it worth the effort since it makes the drawing less ambiguous and increases the overall legibility of the text.
In this example, the erased text is untypically well preserved and largely undisturbed by secondary text which is not the case on many of the other papyri. Work on other manuscripts has already shown that the described method, especially the ‘inking’, needs certain modifications depending on the state of the preserved ink. In cases where the ink remains are very scattered and there are barely any compete signs left, for instance, using the Pattern Stamp Tool is more hindrance than help. The first recording and the sketching process could probably be made easier with more up-to-date hardware or with more finely tuned software settings which would hopefully improve and speed up my drawings in the future.
As a final note, I would also like to express my heartfelt thanks to Krisztián Vértes and the organizers of the workshop ‘Digitale Epigraphik in der Ägyptologie’ in Basel’ for inspiring and helping me develop this technique. Many thanks go to the Museo Egizio for letting me work on their papyri as well as to all my colleagues in Turin, Basel, and Liège for their support.
 for detailed information on the provenance of the Turin fragments, see R. Demarée, K. Gabler, and S. Polis, 'A Family Affair in the Community of Deir el-Medina. Gossip Girls in two 19th Dynasty Letters‘, in Ägyptologische ‚Binsen‘-Weisheiten, IV (in press).
 For a detailed project description see S. Polis, K. Gabler, C. Greco, E. Hertel, A. Loprieno, M. Müller, R. Pietri, N. Sojic, S. Töpfer, and S. Unter, ‘Crossing Boundaries: Understanding Complex Scribal Practices in Ancient Egypt (with a 2019 Progress Report)’, Rivista del Museo Egizio 4 (2020).
 Via the online database Turin Online Papyrus Platform, short TPOP.
 S. Töpfer, ‘The Turin Papyrus Online Platform (TPOP): An Introduction’, Rivista del Museo Egizio 2 (2018)
 e.g. the Virtual Light Table for digitally joining fragments, but also investigating the possibility of automatically finding matching fragments, see S. M. Unter, ‘Transforming Fragments into Documents: Hieratic Papyri and the Use of Machine Learning’, in S. Bickel, K. Gabler, E. Hemauer, and A. Verbovsek (eds), Formen kultureller Dynamik: Impuls – Progression – Transformation (Beiträge des 10. Basler und Berliner Arbeitskreises Junge Aegyptologie (BAJA 10); Wiesbaden, forthcoming 2021).
 See e.g. R. A. Caminos, ‘Some Comments on the Reuse of Papyrus’, in M. L. Bierbrier (ed.), Papyrus: Structure and Usage (London, 1986), 43–61 and C. Eyre, The Use of Documents in Pharaonic Egypt (Oxford, 2013), 33-35.
 This papyrus is currently under study by Martina Landrino (Leipzig University) as part of her PhD thesis, who I would like to thank very much for letting me use this manuscript. For information about the papyrus, text content, and previous publications see http://papyri.museoegizio.it/o/162 (forthcoming). The Museo Egizio di Torino holds the copyright to the scans of the papyrus used in this paper.