Of the several dialects of Coptic that developed in late antiquity, the Sahidic dialect is considered the early classical dialect. Much of the surviving Coptic literature in Sahidic comes from one important late antique repository: the White Monastery in Egypt. One of the most important Egyptian monasteries of the fourth through 12th centuries, it is also known as the Monastery of Shenoute, named after the monk who was the father, or abbot, of the community from the 380s until his death in 465. During Shenoute’s life, the large basilica (the community’s church building) was constructed; over the centuries it was damaged, restored, and changed, but the basic design and construction are from Shenoute’s tenure in the fifth century.

This community is often called the White Monastery, because of the color of the stone used to build the church. Some of the blocks used to construct the basilica were taken from the nearby pagan temple of Repyt (Triphis). Shenoute is the most famous and most important leader of this monastic community, propelling it to a position as a political and cultural center in Upper Egypt.

This monastery’s scriptorium and library were arguably the most influential in the region. Important copies of biblical books and monastic texts written or translated into Coptic have survived [Orlandi 2002] [Emmel and Römer 2008]. Shenoute received letters from the bishop of Alexandria, which were translated into Coptic and circulated throughout the area. Shenoute himself is our most important and probably most rhetorically sophisticated Coptic author [Shisha-Halevy 1986]. And he is particularly known for his stark, prophetic rhetoric in which he condemns sinners, heretics, and others for their failures and predicts the coming of God’s wrath upon them [Schroeder 2006] [Brakke 2007]. The White Monastery, therefore, is one of our most important repositories of Coptic literary manuscripts, and its corpus provides important insights into the religious history of Christian Egypt.

For linguists and Egyptologists, Coptic’s significance lies in its position as the last phase of the Egyptian language family. Egyptian evolved over thousands of years from the third and fourth millennia BCE through the Byzantine era, encompassing Old and Middle Egyptian hieroglyphs of the pharaonic periods as well as Coptic. Because of this connection to ancient Egyptian hieroglyphs, Egyptologists, including Jean-François Champollion, used their knowledge of Coptic to translate the hieroglyphs after the discovery of the Rosetta Stone (e.g., see [Champollion 1824]; [Hamilton 2006]; [Robinson 2012]). Coptic is written primarily in the Greek alphabet, with some modified Demotic characters. (Demotic is the Egyptian script used increasingly during the Hellenistic period, and it preceded Coptic as a written language in Egypt.) The language, therefore, could be loosely understood by the non-specialist as a language of transliteration: Egyptian grammar and vocabulary written primarily in the Greek alphabet (with some native Egyptian letters-Demotic-included). Figure 3 is a transcription of the detail of the Coptic manuscript in Figure 2. The letters rendered in red in Figure 4 derive from the Demotic (Egyptian) alphabet. Some Greek (and to a lesser extent Latin) vocabulary words were incorporated into the language, and after the Arab conquest, some Arabic loanwords came into the language, as well. The Greek loan word akathartos appears in blue in Figure 3. A digital Coptic language corpus would be a resource for research by linguists and Egyptologists alike.

Coptic is also important for Biblical studies, including extra-canonical texts. Many biblical manuscripts survive in various Coptic dialects, including important witnesses in Sahidic from the White Monastery. Shenoute of Atripe is also one of our earliest Coptic authors to cite and quote the bible, providing early evidence as to how the bible was used and interpreted in fourth and fifth century Egypt. Some of our most important extra-canonical texts, including the so-called “Gnostic” library from Nag Hammadi, survive in Coptic. The Nag Hammadi corpus includes philosophical treatises but also a number of extra-canonical Gospels, apocalypses, and other early Christian texts that are fundamental to our understanding of Christian origins. The Gospel of Thomas, shown in Figure 4, is one of the most famous of those texts.

Coptic literature documents the beginnings of the monastic movement, as well. Egypt became one of the cradles of early Christian monasticism when in the fourth century, men and women flocked to Egypt to become monks, and of course Egyptian men and women themselves became monks. Perhaps the most famous of these was Anthony the Great, who lived in a cave in the desert cliffs depicted in or in the region of those depicted in Figure 5. Monastic settlements dotted the Nile Valley from the fourth through eighth centuries.

Thus, Coptic literature is an important primary source for multiple academic disciplines, and we haven't even addressed social and economic history. Countless documentary papyri and ostraca survive, some of which are beginning to be digitized by scholars at http://papyri.info. Ostraca are pot sherds on which people wrote letters, receipts, and other documents. We have thousands upon thousands of Coptic ostraca, many of which are undocumented and unpublished, with more being discovered all the time.

The White Monastery is arguably the most important ancient repository for Coptic texts in the Sahidic dialect and more generally for understanding early Christian monasticism. The community remained a literary powerhouse through at least the twelfth century, accumulating, transcribing, and storing a wide variety of texts. Documents from Byzantine Egypt mention people sending requests and even visiting this monastery’s library to obtain copies of various documents. It contains the largest, earliest collection of contemporaneous, non-hagiographical texts documenting a coenobitic monastery. We have letters, monastic rules, treatises and discourses from Shenoute and at least his next two successors, comprising a corpus larger than that of any other fourth or fifth century monastery. These documents date to a period earlier than the Benedictine material in Italy and Europe, and they comprise a less-hagiographical (and thus more historical) source than more well-known documents about Egypt’s “desert fathers” and “desert mothers,” such as saints lives about famous monks, or the Sayings of the Desert Fathers [Gregg 1980] [Veilleux 1980] [Ward 1975] [Wortley 2014]. When the monastery’s library was “discovered” in modernity, its documents transformed the study of the Coptic language and informed our understanding of the entire Egyptian language family [Emmel and Römer 2008].

In the 18th and 19th centuries, when Europe colonized Egypt and Africa, Europe discovered the White Monastery library. By this time, the monastery was barely populated, and Arabic had taken over as the popular language of Egypt. Even the Coptic liturgies and bibles were in Arabic, not Coptic. The Coptic manuscripts, once part of the most important library in the region, were no longer legible to the Egyptian people, and they were dispersed to libraries, museums, and private collections elsewhere, primarily Europe.

The actual history of the dismemberment of the library is not completely known, but we do know that in the 18th c., pieces of this library were on the antiquities market. Whether taken by European traders or offered up by Egyptian Christians at the monastery, we don’t know for sure. But our first record of White Monastery manuscripts leaving Egypt for Europe is their acquisition by the Borgias in Italy in 1778. Most of the last of the manuscripts were found by the French scholar Gaston Maspero in a small room, essentially discarded by people who no longer knew Coptic, and were taken to Paris [Orlandi 2002].

Today, the manuscripts are scattered across the globe. Only a handful remain in Egypt. The largest collections are in Naples, Vienna, Paris, and England. The map in Figure 6 shows only the locations of manuscripts of Shenoute’s writings. Not where all the biblical texts are located, nor liturgical, hagiographical, and other White Monastery texts. Including all the known White Monastery manuscripts would add to the number of modern repositories on this map. Of course, we do not know the number of documents (whether fragments or whole codices) that exist in private collections.

The texts were dispersed page by page, not codex by codex. Thus, the pages of one original codex might now be all over the globe. Figure 7 is a page from a copy of Volume Three of Shenoute’s Canons for monks; many other folios from his corpus are similarly damaged or fragmented. The chart in Figure 8 illustrates the dispersal of the manuscripts of one text — not a codex but one text — known as Abraham Our Father, which is a letter to monks in Volume Three of Shenoute's Canons. Folios survive from a few copies of Canons vol. 3, so this pie chart represents pages from 6 separate codices. Though six repositories are represented in the chart, each codex's pages went to different repositories. And there are gaps; the text does not survive in entirety even after piecing together all the surviving folios.

Less than 50% of Shenoute’s corpus survives. Of the entire library, the percentage is probably similar, but we don’t know for sure.

One challenge with this level of corpus dismemberment is access: how can we access these texts and understand them with both depth and breadth? What will it require to develop and curate a Coptic corpus suitable for cross-disciplinary, digital and computational research? Some of the challenges to developing a cross-disciplinary digital Coptic research environment include:

Despite the situation of the manuscripts, the late nineteenth and early twentieth centuries saw flurries in the publication of Coptic literary texts: biblical texts, hagiography, liturgy, monastic texts, and others. Issues of access, however, endure. Many texts remain unpublished in whole or in part, and even more are untranslated. In the case of Shenoute’s writings, it was not even until 1993 that a scholar pieced together all the dispersed manuscripts in a codicological reconstruction, enabling various scholars to read all the known surviving pages of any one given text by our most important Coptic author outside of the bible [Emmel 2004].

Copyright and editorial issues also exist. The publication status of Coptic texts is a patchwork, with some texts partially published and partially unpublished due to the dispersal of their original manuscripts in multiple repositories. Some publications are now in the public domain, while others are not. Often, only one publication of any given manuscript exists. (And by manuscript, we refer to the segmented pieces of codices, not entire codices.) Some older, public domain editions are also regarded as problematic by current scholars.

Scholars working in Greek and Latin face similar obstacles in digitizing sources, in that many desirable editions are under copyright. However, in Classics, often there are multiple published editions of any given text, with at least one of them being in the public domain. The Perseus Digital Library (http://www.perseus.tufts.edu) publishes online previous print, public domain editions. This model works for Classical texts but is not sufficient for Coptic literature, because of the different nature of the sources.

Additionally, since the manuscripts themselves have been dismembered, scholars cannot rely upon a given repository and community of librarians to digitize a vital text.

Going back to the original manuscripts for digitization is desirable for basic reasons of scholarship and to provide a digital-native text that is not created from published editions still under copyright. Digitizing the manuscript and encoding all the line breaks, column breaks, page breaks, punctuation, diacritics, and other markings enables the study of many aspects of the language and literature. But it is also time consuming, especially when encoding texts that have not been studied thoroughly, or perhaps have never been translated into a modern language, or even published.

Some existing digital Coptic resources include:

A few resources are available commercially.

The field currently has no generally accepted set of data curation standards for metadata or textual data for Coptic literature. Stable, field-recognized identifiers and digital citation methods for Coptic literature still do not exist. As Joel Kalvesmaki has argued, canonical referencing is complex for text collections with such complicated histories and provenance issues [Kalvesmaki 2014]. The Trismegistos portal provides unique identifiers for various ancient text-bearing objects, but more work remains to be done [Trismegistos]. Existing publications do not even use a common word segmentation practice for printing Coptic text. Only open, collaborative work among various scholars and across digital Coptic projects will ensure that each digital Coptic project does not develop its own unique protocols and datasets that are not applicable for multiple research questions, or that are not readable or transferrable across projects.

Martin Mueller’s keynote lecture at the 2012 Chicago Colloquium on Digital Humanities and Computer Science noted philology’s tendency toward perfectionism [Mueller 2012]. Philologists' resistance to releasing editions “too soon” bumps up against the digital scholarly world’s impulses to release data openly and swiftly, in order to increase access to texts. The ability to edit and annotate a digital corpus resolves some of these anxieties, provided that the technology of the environment for the corpus enables revisions. An example of such a corpus is the Papyri.info site, where the Papyrological Editor enables further edits and refinements.

The recent controversy over the so-called Gospel of Jesus’s Wife Fragment demonstrates the necessity for an open, well-curated, and annotated digital Coptic corpus. Harvard University announced the discovery of a papyrus fragment containing a small portion of a Coptic literary text containing the terms “Mary,” “Jesus,” and “my wife” in September 2012, and a slew of media fanfare followed, including a front-page story in the New York Times [Goodstein 2012]. The primary questions were: is this document an authentic ancient text and does it provide evidence that Jesus was married — possibly to Mary Magdalene? — or at least that a particular early Christian group wrote about Jesus marrying Mary Magdalene [Le Donne 2013]? For two years, scholars debated the authenticity of the fragment, examining the vocabulary, syntax, and handwriting. Similarities to vocabulary in the Nag Hammadi Library were noted, although the grammar of the Coptic in the fragment struck some scholars as inauthentic very early in controversy [Robinson and Halton 2012] [Čéplö 2012]. And since the fragment appeared on a small piece of papyrus, it seemed possible the text could be an amulet, spell, or prayer fragment.

Immediately, the potential for analyzing the document against digital corpora curated for Coptic’s particular language structure and annotated linguistically became apparent. The lack of such corpora meant that only one scholar made such an attempt, using a widely circulating digital version of the Sahidic New Testament gospels and the Nag Hammadi Library’s Gospel of Thomas [Čéplö 2012]. These digitized texts, however, however, contain many irregularities in spelling and word segmentation, and are not annotated. While helpful, they could not be called well-curated nor comprehensive. Scholarly debate over the authenticity of the fragment continued, focusing on traditional methodologies of studying ink, papyrus, handwriting, and grammar, in blogs, social media, and traditional scholarly outlets [Goodacre 2014b] [Goodacre 2014a] [King 2014a] [Choat 2014] [Yardley and Hagadorn 2014] [Azzarelli, Goods, and Swagger 2014] [Hodgins 2014] [Tuross 2014] [Depuydt 2014] [King 2014b]. Finally Coptic scholar Christian Askeland determined the text was indeed a modern forgery, since another document using the same ink and written in the same hand was also forged. He published his results in a blog post and later a traditional article [Askeland 2015]. Traditional scholarly methodologies informed this debate, but open access corpora and images would have accelerated the findings and would have allowed for testing of hypotheses speculated between 2012 and 2014.

A richly annotated corpus of digital Coptic literary texts in an open-access environment, with well curated metadata, and which adheres to existing digital and traditional field-based standards, enables the exploration of the multidisciplinary research areas we have described. Such work requires tools to process and annotate Coptic text, a search and visualization infrastructure, and community-based standards. Coptic SCRIPTORIUM is developing such tools and technologies, plus a digitized corpus created with these tools.

To create a digital corpus of Coptic texts suitable to automated search and other digital and computational methods, new technologies must be developed specifically for processing the Coptic language, and existing technologies must be adapted. Such technologies include:

There is no universal, uniform method of word segmentation in Coptic; different scholars have followed different guidelines. The emerging standard in the United States is currently that of Bentley Layton, in his Coptic Grammar [Layton 2011]. In Germany, however, some scholars follow the paradigm established by Walter Till [Till 1960]. Tokenizing Coptic is essential for linguistic study, since each morpheme needs its own part-of-speech annotation. Historical research using basic vocabulary searches also requires tokenization. Finally, more complex research into style and rhetoric, authorship attribution of unidentified texts, and searching for text reuse and quotation (especially using algorithms to search for biblical citations) demand a tokenized text

Other existing, community standards used in both digital and print scholarship can be adapted, as well. Trismegistos numbers from the Trismegistos database of ancient texts and texts-bearing objects should be included in metadata to enable linked data across digital projects [Trismegistos]. For works specifically in Coptic, the CMCL abbreviations for manuscript codices (e.g., MONB.YA = White Monastery manuscript YA) and Clavis Coptica for authors and texts are existing standards [Clavis Patrum Copticorum] [Suciu 2012]. For texts by Shenoute, the incipits and abbreviations for texts developed by Stephen Emmel in Shenoute's Literary Corpus should be used for metadata [Emmel 2004]. The field still needs a system for Uniform Resource Names (URNs) and other more refined citation methods for curating digital data and documents — especially given the dismembered nature of the corpus, and that some pieces of the corpus may be published in born-digital formats while others may be adapted from previously published editions.

Tools to annotate or mark up the digitized, curated text enable search and computational research methods for work in a variety of disciplines need to be developed. We know of no other open source tools to annotate digital texts in the Egyptian language family. SCRIPTORIUM is working on the following technologies

At Coptic SCRIPTORIUM textual annotations are made using multi-layer and standoff markup [Carletta, Evert, Heid, Kilgour, Robertson, and Voorman 2003] [Dipper 2005], which can capture the variety of annotations: linguistic, paleographic, philological, etc. The token layer is the base layer of data, the smallest unit of data annotated, as seen in Figure 11, which shows the annotation of two bound groups (one in red, one in blue), which translate into “of a son of Abraham”:

The layers provide data for the morphemes, the bound group, the line number in which the text appears in the original manuscript (lb, following the TEI/EpiDoc standard), the column in which the text appears in the original manuscript (cb), the page on which the text appears in the original manuscript (pb, in which MONB.YA is the siglum for White Monastery codex YA as designated by the CMCL standards), the normalized morphemes, and the part of speech tags for the normalized morphemes (PREP=preposition, ART=article, N=noun, NPROP=proper noun).

SCRIPTORIUM provides the means to search the multiple layers of data in various combinations, including in conjunction with metadata. We use the open-source search and visualization tool ANNIS [Zeldes, Ritz, Lüdeling, and Chiarcos 2009]. ANNIS contains built-in visualization capabilities and can be customized for each corpus. Coptic SCRIPTORIUM has embedded a Coptic keyboard, a web font for Coptic Unicode characters, and various visualizations of the data. All the tools and corpora discussed in this essay are available at http://www.copticscriptorium.org.

The multilayer architecture allows for multidisciplinary research [Krause and Zeldes 2014]. Historians may be interested in vocabulary searches on the normalized words. Linguists may query the parts of speech for computational morphological and syntactic research. Scholars working on ancient prosopography or network analysis may search for named entities. Additional layers may be added if other researchers wish to annotate the corpora for other research questions. Since all data documents are licensed with Creative Commons Attribution licenses, philologists and paleographers who wish to publish their own digital editions of manuscripts may download, modify, and annotate or re-annotate our XML documents for their own work as long as they provide attribution to the source.

The community of Coptic scholars is small, but the impact of our work ripples out into many fields. The true hope for digital scholarship in the Coptic language and literature lies beyond our individual efforts and in the community of Coptic scholars within and outside the academy: scholars who digitize texts, write annotations, inspire and develop new technologies, conduct research using the platform, and contribute to the evolving standards.