Perseus Digital Library Updates

Gregory Crane
Professor of Classics
Tufts University
September 3, 2019

I am writing to report that the Perseus Digital Library had the honor of receiving support for an NEH Digital Humanities Level 3 Project: “Beyond Translation — new possibilities for reading in a digital age.” While this is just one project with limited funding, it reflects a larger potential shift for the study of Ancient Greek and other languages. When I began my career as a graduate student, more than a generation ago, specialists in languages such as Ancient Greek could only direct full scholarship at other specialists. Now, however, we are in a position to frame our understanding of such languages in a form that makes sources immediately accessible to non-specialists. From my perspective, this reflects a fundamental shift in the audience and the realizable goals for those of us privileged to earn a living as specialists on earlier languages from the human record.

The basic premise of the funded project is simple: the rapid evolution of reading environments has begun to open up a third path for reading, one situated between mastery of a language and dependence upon translations. For me personally, the need for such a third path weighs on me every time I log into Netflix and confront offerings not only in French, German, (various forms of) Spanish, Arabic, Mandarin and other widely taught languages but also in languages such as Turkish, Korean, Malaysian, and Hindi. Even if I had access to classes in each one of these languages, I would never have time to master them — and there is always another language. 

Of course, this programming is (for the most part) designed for an international audience and this entails a certain cultural leveling: while Netflix needs content that appeals to local audiences, it needs content that will also engage audiences with many different linguistic and cultural backgrounds. It would be interesting to explore the transnational elements to which these local productions appeal (and this is, of course, an academic enterprise in itself) but for me at least I find myself constantly drawn to inevitably local context. When I watch the Protector, I find myself wondering how I might, if I had grown up and been educated in Turkey, understand explicit references to Ottoman history and the implicit associations with particular scenes? What would I understand about the historical Korean landscape that provides the setting for the Zombie series Kingdom? What elements of class and cultural background are lost as I follow the subtitles of Diablero, a kind of dark “Ghostbusters” set in Mexico City? My training as a philologist makes me ache to push beyond the subtitles that flash across my screen but I feel only discomfort and a kind of vertigo as I scroll the offerings online. And, of course, even though Netflix may include content produced in far more languages than any normal human being can master, these will nevertheless constitute a relative handful of languages with large potential audiences. If I move to a venue such as YouTube, the range of content grows even more broad and disorienting.

There is, of course, an entire industry centered around localization. As my wife and I watch Americans decide where to live on International House Hunters, we don’t think about the fact that the Discovery Channel itself aims for a global audience and, reportedly, translates more than 100,000 hours of content a year. A&E translates into more than 20 different languages. On the industry forum Slator.com, the driving question is how far rapidly evolving machine translation can augment — and even replace — the work of human translators. Given the scale at which new content is produced and the dizzying range of languages spoken by potential audiences, the pressure for automation is immense and will surely drive technical improvement. I may well find my way to an industry meeting on localization in Amsterdam at the end of November. After all, what else should a philologist with a particular focus on Ancient Greek do?

The preceding rhetorical question is not facetious.  When vertigo takes hold as the range of languages and cultural contexts available in modern media begins to overwhelm me, I turn back to Ancient Greek and, for now, to the Homeric Epics. The reason why I feel such anxiety when I view cultural commodities from around the world is because I have a deep sense of how limited the best translations can be. Languages do not frame the world in the same way and our cultural linguistic backgrounds shape how our experiences affect us. When I turn to Homer, I see an optimal space within which to experiment with ways audiences can push beyond translations to engage with a complex cultural product from a profoundly alien context.

• The vast majority of readers who engage with the Iliad and Odyssey do so — and will always do so — through translations in languages that they understand (ideally their native languages).
• Where the immediate audience for series available in venues such as Netflix may be limited, the half-life of those audiences is uncertain — is it worth spending a decade to develop intellectual infrastructure for a series for which the immediate appeal may grow rapidly dated? The Homeric Epics may never have mass appeal (although Emily Wilson’s translation of the Odyssey provides an exception to that within the anglophone world), but poems that have commanded human attention and emotions for centuries, if not millennia, have a track record that suggests that they will continue to engage audiences in the future. I see this every year when I have a chance to introduce these poems to a new cohort of students. 
• A rich body of information about the Iliad and Odyssey is available in various digital forms in the public domain and/or under open licenses. This includes specialized lexica, commentaries, at least one critical edition of the Greek text and one diplomatic transcription of a major manuscript (with ancient commentary), multiple translations in multiple modern languages, grammars of Homeric Greek, and explicit analyses of the morphological and syntactic function of every word in each epic.
• A dynamic, collaborative and international community of Digital Classicists has emerged that contributes data, software, and ideas of all kinds.

We are in a position to develop a new mode of reading, one that challenges audiences to see in translation a starting point rather than an end. What is the relationship between the literary translation and the Odyssey or between the subtitles on your screen and the words you are hearing? What are you missing? How much can you understand on your own? These are the questions that we can ask as we work with ancient Greek? A century ago, a handful of students had an opportunity to study Ancient Greek and Latin from childhood and to benefit from an academic infrastructure that offered extended linguistic training. In the twenty-first century, virtually no one in the anglophone world (at least) has access to such training. From my perspective, as my head spins at the linguistic variety of online content, I see the lack of extensive training as an opportunity. The grand challenge before us in a complex world is not (simply) to master one or more languages but to understand how — and, of course, how far — we can engage with the languages that we can never learn and the cultures in which we can never immerse ourselves. 

Of course, we are at the very beginning of a larger process. We need to absorb questions and ideas from a wide range of disciplines. We need to engage with others not just in the usual established universities but with colleagues around the world and with points of view from beyond academia. But we have an opportunity to rethink — and rearticulate — the value of what we learn from studying languages such as Ancient Greek (or Latin or Coptic or Classical Chinese or Sanskirit or Classical Persian or any historical language). Every time I turn back to Netflix and think of industrial localization, I catch glimpses of what we can contribute to our students and to the world beyond academia.

Returning now to the NEH project, the proposal opens with the following paragraph: “Our goal in this project is to promote a fundamental change in how human beings view translations and the cultures of which their original source text is a product. In a period where marketers and the popular press focus on artificial intelligence, machine translation has become increasingly important. But for us as humanists in general (and as philologists, focused on understanding the human record as deeply and broadly as possible), translation is only one instrument among many and, if seen as a final goal, a barrier to understanding others whose parents spoke to them in languages other than our own and who look back upon cultural traditions very different from those of most Americans. At some level translation obscures more than it reveals, offering a deceptively simple equivalent. Even when we translate between European Spanish and North American English — languages that evolved together within Europe —  research shows that terms such as “shame” and “vergüenza” mean fundamentally different things to speakers in the United States and Spain. Our goal is to allow readers to push past the translation and to explore for themselves how original sources frame human experience.” 

I will end by emphasizing the mix of continuity and innovation that I personally see in this project.

• For me this project began in fall 1975 when, as a first year student in college, I discovered that Gregory Nagy was having his students learn the Greek alphabet, look up forms of a word in the print concordance and then compare the Greek to the Richard Lattimore English translations of the Iliad and Odyssey (which followed the Greek line numbers). I was one of the lucky students who had had a lot of Greek before college and I have never forgotten how surprised I was at how much students, with no knowledge of Greek, learned with this (now primitive) combination of print resources. I never forgot the impression that that made upon me and that impression shaped my earliest work with Perseus. Already in the late 1980s, students were using side by side translations in the Hypercard version of Perseus to explore the meanings of Greek words on their own. I have had the privilege of working on the Open Greek and Latin Project  (http://www.opengreekandlatin.org/) with Greg Nagy, Lenny Muellner and colleagues at Harvard’s Center for Hellenic Studies. The Center is now partnering with us on Beyond Translation. We now have an opportunity to make what was possible for Homer because of the concordance and Lattimore translations become a core component for ambitious readers working with a far wider range of sources.

This project builds upon preliminary digital work by many people over many years. In particular, we are extending work that David Bamman, now an Assistant Professor at the School of Information at Berkeley, began when we had the privilege of working with him at Perseus a decade ago. The proposal that we submitted in 2019 allowed us to carry forward the work that David began but it also depended in large measure upon work done by Sophia Sklaviadis, in fall 2018, her first semester after joining the Computer Science PhD Program at Tufts. She came to Tufts to learn how to apply language technologies to Ancient Greek and she, like David Bamman (and David A. Smith and David Mimno) represents a new role that did not exist when I began work on what we now call Digital Classics: Computer Scientists who are committed to, and deeply conversant with, issues in the humanities. The funded proposal also built directly upon work on the Scaife Digital Library Viewer (https://scaife.perseus.org/) developed by James Tauber, CEO of the Eldarion Web Application Development company and long-time student and supporter of Biblical Greek. While we have been able to find support to help James pay his bills and his employees, his contributions went far beyond the contracts that we were able to offer. Here again, we find a new role for contribution that I don’t think existed or could have existed when I began work with Digital Classics in the 1980s. A century from now, students of Ancient Greek and other historical languages will, I believe, see work by dedicated researchers from the computational sciences as among the most important — perhaps essential — contributions.

Gregory Crane
Tufts University — Fall 2019
Greek 1: Introduction to Ancient Greek and the Homeric Iliad
MWF: 9:00-10:15
Eaton Hall 209

Are you interested in studying a new language in fall 2019? Whether you want to try something different for your language requirement or you have a year — or even a semester — you have an opportunity to travel thousands of years into the past and to confront the oldest sources in the continuous tradition of European literature.This is not your parents’ language class, and it is not high school Latin. You have an opportunity to participate in the reinvention of an ancient field and the development of a new track within the humanities as a whole. And you also have a chance to begin developing a research agenda of your own, one that can bring together the humanities and emerging fields such as data science. 

No language is poised to benefit more profoundly from disruptive technologies than Ancient Greek. This disruption is fundamentally changing the ways in which we can interact with this, and other ancient languages. And while the changes can enhance the questions that specialists can pose, those who are just starting to learn Ancient Greek are poised to benefit the most. A growing range of reading support tools allow students of the language to interact with the primary sources immediately. The more understanding of the language and the culture students internalize, the richer the experience will, of course, be — there is a big payoff to sustained study. But the tools now becoming available mean that even a semester of study would allow students to engage with sources in Greek directly, at a depth that would have been impossible before.

Disruptive new reading tools challenge us to rethink our understanding of historical languages and cultures such as those of Ancient Greece. None of the scholarly infrastructure that has emerged over generations of modern academic study, over centuries of print culture, and over millennia of sustained scholarship has been organized to match the opportunities and challenges of an interactive reading environment that exploits computational methods in general and machine learning in particular to serve a global audience (not just those familiar with the languages of European scholarship). 

The creative disruption opens up a wide range of opportunities, especially for ambitious students who wish to contribute and to develop a research portfolio. Participation in the 2018/2019 introductory Greek class at Tufts University enabled three students to begin research projects of their own. Two, Madeleine Harris (Tufts ‘21) and Pearl Spear (Tufts ‘22) applied emerging research from cognitive science and natural language processing to begin transforming our understanding of ancient emotions, while Bella Hwang (Tufts ‘22) combined what she learned in Greek and in her courses on computer science to begin modelling the linguistic features that students need to master if they are to read Homeric Greek with fluency. Such projects allow students to make contributions to the study of Ancient Greek, to spend extensive time posing new questions of poems that have fascinated humans for thousands of years, and to begin cultivating the most dynamic analytical methods of the 21st century. 

The 2019/2020 academic year will be the second step in a multi-year process re-engineering the way our students engage with and master Ancient Greek. We will focus on the Homeric Iliad — the earliest surviving literary source in the continuous tradition of European literature. No secular literary work from Europe has commanded the attention of more audiences, from more cultural contexts, over more time than the Iliad. As you begin to study Homeric Greek, you travel back thousands of years, hear the voices of oral poets who worked with a tradition that was already thousands of years old and engage with a world that is both alien and familiar.

Call for applications: Digital Editions in Practice, A Two-Day Workshop

The Perseus Digital Library at Tufts University will host a two-day workshop that provides an overview of a sample, practical digital editions creation workflow. This will feature both an open-lecture component led by developers and expert users of advanced technologies and “hands-on” sessions for participants that offer in-depth demonstrations of select tools and technologies as well as discussions tailored to the attendees.

For the full announcement, please see: https://goo.gl/BGbWxy

Gregory Crane
May 1, 2018

At the end of March 2018, my collaborators and I finished enjoying five years of support — 5,000,000 EUR(!) — from an Alexander von Humboldt Professorship, support which allowed young researchers from many different countries to work both as a team and on their own. Documenting all that work will be a significant task and requires its own publication(s). Work, at Leipzig, Tufts and elsewhere, on Open Greek and Latin (OGL) and on the Canonical Text Services (CTS) protocol upon which OGL builds provides the starting point for much of the work described below. A tremendous amount of support for OGL came from the European Social Fund and the Alexander von Humboldt Foundation, but collaborators at Perseus at Tufts University, at Mount Allison University in Canada, at the University of Virginia, at the Harvard Library and Harvard’s Center for Hellenic Studies (CHS) have contributed time and significant sums as well. As a group, they have made 37 million words of Greek and Latin available in CTS-compliant epiDoc TEI XML via GitHub.

This paper, however, does not focus primarily upon what happened at Leipzig but takes note of a number of events that have taken place in the opening months of 2018 and that have some connection to, but also depend upon efforts outside of, the Digital Humanities Chair at Leipzig. Each taken separately is important. All of these events taken together reflect a broader, systematic change — and change for the better — in Ancient Greek and Latin philology in particular and, ultimately, for all philology.

For the full article, see https://goo.gl/zG4yT4.

On March 15, Eldarion released the initial version of the Scaife Digital Library Viewer. The release is, of course, a first step, but this first step changes the world in at least two fundamental ways: (1) Perseus is alive — it can finally include new materials on an on-going basis; (2) the Scaife Digital Library Viewer provides a foundation for an environment that can publish a growing range of born-digital, openly licensed, and networked (and fully networkable because they are openly licensed) annotations and micro-publications that cannot be represented in the incunabular digital publication systems that still internalize the limitations of print publication.

First, Perseus can now be configured so that it can include new materials almost immediately. We have not yet established a regular workflow — the initial Scaife Digital Library Viewer still runs on a server maintained by Eldarion rather than Tufts — but updates on a weekly and even a daily basis, if not real time, would be quite reasonable. New content does not even have to be in Greek or Latin — we already include a Persian edition of the Divan of Hafez. More importantly, if someone outside of the extended network of Perseus collaborators puts their content in the right format (for now CapiTainS-compliant EpiDoc TEI XML), we can include it. Thus, Neven Jovanovic was able to publish the first of what is expected to be a series of early modern Latin texts in Perseus (Scaliger’s Latin translation of Sophocles’ Ajax). Prof. Hayim Lapin from the University of Maryland converted his CC-licensed version of the Hebrew Old Testament , Talmud, and Mishnah. At present, anything that ends up as visible in the Scaife DL can be (because we require an open license) a permanent part of the Perseus collections. We need to think through a general process of content submission (and however open we wish to be, there are obviously some limits), but there are enough established collaborators with content to add and enough CC-licensed material that we would like to add that we already have enough materials to test a workflow for updates.

Second, the use cases of Perseus and of Digital Classics are not only more varied than those of print but involve so many data types and so many implicit use cases that they represent an emergent system. These include born-digital critical editions (with variants classified and dynamically configurable), diplomatic editions with alignments between transcription and source images, alignments between different versions of the same text in the same language, bilingual alignments between source texts and translations morphological and syntactic analyses, co-reference resolution, and other categories of linguistic annotation, social networks, geospatial data, representations of digital intertextuality (including annotations expressing estimating probabilities that a given word or phrase represents a paraphrase or direct quotation from a lost source text), and an unbounded set of potential new annotation classes. Use cases include not only specialists posing new kinds of questions (e.g., search a corpus for instances of “future less vivid conditionals” or a semantically clustered list of verbs associated with male vs. female agents) but a fundamentally new mode of interaction that we might term language wrangling or language hacking, where readers have such dense networks of explanatory annotations that they can engage immediately, at some level of precision, with any annotated source in any language, whether or not they have any prior of knowledge of that language. Such reading is a new form of engagement that lies between the experience of experts who have spent their 10,000+ hours immersed in a subject and the passivity that a print modern language translation, with no mechanisms to get past its surface and into the source text, imposes upon the reading mind.

Looking at the first release of the Scaife Digital Library Viewer, it is easy to see all the work that needs to be done. Indeed, for me, the steady progress towards a Perseus 5.0 only deepens my appreciation for what went into the development of Perseus 4.0 (the Java-based version, initially developed by David Mimno more than fifteen years ago and still in use at www.perseus.tufts.edu) and Perseus 3.0 (the Perl-based version that David A. Smith initially developed on the side to give Perseus its first web presence back in 1995). More than a decade ago, we solved another, less immediately obvious problem for having Perseus emerge as a place in which to publish content. In March 2006 (after being badgered by Ross Scaife, as well as Chris Blackwell, Gabby Bodard, Tom Elliott, Neel Smith and others), we began to apply a Creative Commons license to content that had no legal entailments. As soon as we decided that we would create collections that only contained CC-licensed content, we solved the bottleneck problem: so long as we actually made the content available, we could never use exclusive control over that content to restrict the development of services that we could not provide (say hello to Perseus Philologic, Alpheios.net).

The Scaife Viewer of March 2018 may only be a beginning. It may have a great deal more to do (e.g., integrating treebanks and source text/translation alignments). But the code is open and the possibilities are almost unbounded.

Gregory Crane
March 15, 2018
Leipzig, Germany

I am pleased to announce the first release of the Scaife Digital Library Viewer, a reading environment for source texts that follow the Canonical Text Services (CTS) data model. Our initial focus is on pre-modern sources, but the underlying approach applies to source texts of all periods. CTS provides a framework within which we can cite particular words in particular versions of particular texts — whether a version is a papyrus, manuscript, or a critical edition, whether versions of that text derive from a single lost original (as is the case for many ancient Greek and Latin texts) or the text itself appears in many versions, each of which has comparable authority (as is the case for many medieval sources). For those interested in more information, James Tauber, lead developer for this release, will present the Scaife Digital Library Viewer online on April 26 at 5 pm CEST as part of Sunoikisis Digital Classics. The presentation will be recorded and available, along with any other course materials, on the SunoikisisDC website after the class itself.

The Scaife Digital Library builds upon the Capitains suite of tools for creating and managing CTS-compliant textual data, developed by Bridget Almas, then one of the two leaders of the Perseids Project, and now at the Alpheios Project (http://alpheios.net/), and by Thibault Clérice, then a member of the Alexander von Humboldt Chair of Digital Humanities at Leipzig (and now at the École nationale des chartes). James Tauber, leader of Eldarion, a web development company as well as a long-time student of, and developer for, Biblical Greek, oversaw the development of the Scaife Digital Library as an open-source, customizable reading environment. In memory of Ross and what he stood for, the release is intended to empower the community to take charge and carry work forward. And, of course, the code is open and available on Github. Ross would not have had it any other way.

Ross Scaife (1960-2008) was a pioneer in reinventing the study of Greco-Roman culture to exploit the possibilities of a digital age. He was among the first — if not the first — to get tenure for a purely digital publication, Diotima: Materials for the Study of Women and Gender in the Ancient World, a project that he and Suzanne Bonefas launched in 1995 (the same year that David A. Smith, now an Assistant Professor of Computer Science at Northeastern, established the first web presence for Perseus). Ross was a colleague and he was a friend, whom we mourn still and will always miss. We lost him on March 15, 2008 and it is with fond memory that we announce the first version of the Scaife Digital Library in his honor, on March 15, 2018, ten years later.

Gregory Crane
February 24, 2018

Who is using Clarin (https://www.clarin.eu/) and/or Dariah (https://www.dariah.eu/), and particularly the German subprojects https://www.clarin-d.net/de/ and https://de.dariah.eu/, to work with historical languages?

If so, are you doing so as a funded member of Clarin or Dariah?

Who has used the Dariah repository (https://de.dariah.eu/repository) to store data? I have found documentation about how to add data but I have not yet found any collections stored within the Dariah repository?

What other services offered by Clariah and/or Dariah have you used? Are you planning to use any?

Ideally, this would generate a public discussion in forums such as Twitter, Humanist and the Digital Classicist mailing list but feel free to email me directly on gmail (gcrane2008).

The Perseus Digital Library at Tufts University invites applications to “Digital Editions, Digital Corpora, and new Possibilities for the Humanities in the Academy and Beyond” a two-week NEH Institute for Advanced Technology in the Digital Humanities (July 16-27, 2018).  This institute will provide participants the opportunity to spend two intensive weeks learning about a range of advanced new methods for annotating textual sources including but not limited to Canonical Text Service Protocols, linguistic and other forms of textual annotation and named entity analysis.  By the end of the institute, participants will have concrete experience applying all of these techniques not just to provided texts and corpora but to their own source material as well.

Faculty, graduate students, and library professionals are all encouraged to apply and international participants are welcome. Applications are due by February 1, 2018.  In order to apply for the institute, applicants need to 1) complete the online registration form; 2) concurrently send a CV and statement of purpose  by email to perseus_neh@tufts.edu

Full application information regarding the statement of purpose and other important details may be found here: https://sites.tufts.edu/digitaleditions/applications/

For more information please visit the institute website: https://sites.tufts.edu/digitaleditions

This Institute builds upon experiences from, and work subsequent to, “Working with Text in a Digital Age,” a 2012-2014 NEH IATDH project and the on-line seminar, Sunoikisis Digital Classics.

“Deconstructing the Open Greek and Latin Project: The First Thousand Years of Greek”

An AIA-SCS Pre-Meeting Workshop, presented in coordination with the SCS 


January 3, 2018, 9:00 to 5:00, Tufts University, Medford, MA

Interested in open access, the digital humanities, or conducting digital scholarship in your research and/or teaching?  Aren’t sure what these topics have to do with classics or archaeology, or even how to get started?  Then, please consider joining us next January 3 at the AIA-SCS pre-meeting workshop “Deconstructing the Open Greek and Latin Project”!

In this workshop, partners from the Perseus Digital Library, the Harvard Library and Harvard Center for Hellenic Studies, the Institute for the Study of the Ancient World, the University of Leipzig, Mount Allison University, and the University of Virginia Library will come together to demonstrate research tools, explain how to involve students in digital scholarship, provide open data for hands-on exploration from the Open Greek and Latin Project, as well as create a growing and supportive open access community.

Tools and technologies we’ll work with include GitHub, Oxygen, TEI-XML and EpiDoc

Registration is offered on a “first-come first-served” basis and the workshop is offered free-of-charge with a registration deadline of Friday, November, 3, 2017.

To register, please complete our registration form!

For more information please visit the workshop website at http://sites.tufts.edu/oglworkshop or contact us at ogl.workshop@tufts.edu

Presented by the Forum for Classics, Libraries and Scholarly Communication of the Society for Classical Studies. Sponsored by the Perseus Digital Library at Tufts University. Co-sponsored by the The Center for Hellenic Studies; Harvard Library; and the Institute for the Study of the Ancient World at New York University in collaboration with the Humboldt Chair of Digital Humanities at the University of Leipzig; Mount Allison University; and the University of Virginia Library.

Alison Babeu called my attention to a recent blog by a Princeton Classics undergrad that really captured a major challenge and opportunity for a new Perseus. Solveig Lucia Gold described her own reaction to the ups and downs of using the reading support that Perseus has offered for Greek and Latin for decades (and, indeed, since before many of our undergraduates were born, if we consider the CD ROM versions of Perseus). The situation will be even better — or worse — when we finally integrate treebanks and alignments between the source texts and the translations. At that point, you can puzzle out almost any text in any language. We have treebanks (morphological and syntactic analyses) for every single word in the Homeric Epics, for example — you have interpretations for any sentence in these epics.

But you can’t read Plato’s Rebublic or the Iliad or the Diwan of Hafez (to shift to Persian) by looking up every single word — true reading and true appreciation requires that we internalize as much of a language as possible.

The goal is not to replace learning but to provide a scaffolding whereby we can go from no knowledge to as much internalized understanding as we have the time and determination to acquire. If I were to pick one challenge for the coming ten years, it would be to create the framework to foster such learning. (And here I look forward to the next generation of work from Alpheios.)

There is no greater topic for research in historical languages than enhancing the ways in which we human beings are able to learn those languages — a question that is technical, social, and profoundly intellectual, for it challenges us to understand why we care about the past as much as we do — and why we should care even more.