Digitizing the classics

Long before the Italians rediscovered original Greek sources during the Renaissance, Arab scholars recognized the importance of ancient science and philosophy and began translating precious writings into Arabic. Now, Classics Professor Mark Schiefsky wants to transform those ancient Greek texts and their Arabic translations into an open-access digital corpus that could provide important insight into the development of science in the classical world.

During the Abbasid period, which began in the mid-eighth century, Islamic caliphs started sponsoring the translation of ancient Greek and Roman texts. While Arabs had their own literary traditions and did not systematically translate Greek literature, they were interested in Greco-Roman mathematical and medical treatises and philosophical writings.

“People recognized that Greek texts contained a lot of knowledge that superseded the knowledge available in the Arab world at that time, and realized that it would be fruitful to adopt that knowledge,” Schiefsky explained.

He added that the decision to translate these texts was motivated in part by a desire to compete with the Byzantine Empire to the West.

“The Arabs wanted to say they were the true inheritors of the Greek tradition,” he said.

But many ancient texts were also translated for practical reasons. The writings of Galen, a prominent second century physician, had an important influence on medicine in the Arab world, while the Greek philosopher Aristotle’s writings on logic were exploited in disputes over Islamic law. Even today, classical texts continue to resonate in the Arab world, Schiefsky said, citing Iranian Ayatollah Khomeini’s study of Plato’s “Republic” in creating the Iranian state.

Schiefsky recently received a two-year grant from the Mellon Foundation to support the creation of this new, structured corpus of digitized Greek and Arabic texts. The corpus, a collaboration with the Perseus Project at Tufts University, will be used for studying translations of Greek texts and their reception in Arab culture up until the present.

The Greco-Arabic “bilingual lexicon,” as he calls it, is not the first project in which Schiefsky has used sophisticated technological tools to serve humanistic research. The Archimedes Project, which he led in collaboration with the Max Planck Institute for the History of Science in Berlin, assembled myriad scientific texts in different languages, allowing for new investigations into the history of mechanics. As with the Archimedes project, the vastly wider body of information that will be available in the bilingual lexicon will enable researchers to pose new statistical questions about how particular features of texts change over time. The lexicon thus represents a kind of shift from the traditional philological approach, with its focus on words and details, to a more comparative approach.

“How do conceptions of medicine, say, or mathematics, change over the long term when we move from Greek to Arabic to Latin sources?” Schiefsky asked. “To address such questions in a comprehensive way requires taking a huge corpus of material into account. Modern information technology offers many new tools and approaches for such analysis, which are only now beginning to be applied in the humanities on a large scale. Despite a large number of digitization efforts over the years, there is still a lot of work to do just to get the basic data in a suitable form.”

A large body of Greek writings from Homer up to 600 A.D. has already been digitized by the Thesaurus Linguae Graecae, though the thesaurus is not available in the public domain.

Schiefsky’s project will contribute additional Greek texts in areas like science, medicine, and philosophy, as well as Arabic texts that are mostly, but not entirely, translations from the Greek. One of the database’s most important features, he said, will be correlating parallel sections of text, allowing scholars to compare phrases or passages page to page.

“Searching is very nice, and Google is very good at searching. But you can do a lot more than search every time a word appears,” Schiefsky remarked, citing examples like determining how frequently certain terms were used at different points in history.

“I’m interested in the development of knowledge and the development of science, so you need good linguistic tools to do that,” he said.

A member of Harvard’s Digital Humanities Working Group, Schiefsky believes strongly in harnessing open-access technology for the benefit of collaborative scholarship. The digital corpus will be entirely open access, using a Creative Commons license that allows other scholars to use and improve the software.

“We’re moving away from a way of working in the humanities with one scholar making a change to a text that is incorporated into future editions for eternity, and toward more collaborative methods,” Schiefsky said.