Beyond the book
Sefaria’s Lev Israel brings Torah to the digital age
Lev Israel knows he is a key player at a pivotal moment of Jewish history.
As chief data officer for Sefaria.org, he is leading a technical process that already has placed more than 150 million words of Jewish texts online for free use by anyone.
He is aware that Judaism has a tradition of resisting new media. “Matters that were written you may not express orally, and matters that were taught orally you may not express in writing,” the Talmud declared.
And yet the Oral Law — the Mishna and the Talmud — was in fact written down and published. “It is time to act for the Lord,” the Talmud explained.
These Talmud texts are at my fingertips right now because they’re part of the source sheet that Mr. Israel has prepared for his talk, “Old Wine in New Vessels — Torah, Sefaria, and New Media,” which he will give on Sunday in Teaneck. (See box.) You can read it yourself at www.sefaria.org/sheets/15490. It’s one of more than 100,000 online pages that combine Hebrew texts, translation, and the collators’ notes that have been created by Sefaria users. The site offers tools to select excerpts from its library and create pages that can be handed out to students. All told, some 4,400 people have either created source sheets or contributed translations to Sefaria. The site boasts 150,000 users each month.
“This is probably the most interesting analogue to where we are now,” Mr. Israel says. He’s talking about the rabbinic discussions of writing recorded in the Talmud and collected in his source sheets. “The pain and difficulty they felt in moving from an oral culture to a written culture has something of an analogue in how we feel in moving from our print culture to a digital culture.”
Mr. Israel, who grew up in Albany, went to Sefaria four years ago with a background in both yeshiva study and Israel’s high tech world. He had studied computer science and philosophy at Boston University before moving to Israel for 18 years, shuffling between the high tech start-up world and the beit midrash — the yeshiva study hall.
Being in charge of data at Sefaria means that Mr. Israel is both a publisher and a computer scientist. As a publisher, he’s overseeing the copy editing of translations of classic Jewish texts, ranging from the ninth to 20th centuries, that Sefaria commissioned. Only about 10 percent of Sefaria’s library is in English and he wants to expand that. A grant enabled Sefaria to republish the English translation of the Talmud overseen by Rabbi Adin Steinsaltz and being published in print by Koren Publishers. He’s licensed the Jewish Publication Society’s Tanakh translation, and has worked with several other Jewish publishers to put their translations on line.
In general, Mr. Israel said, Sefaria tries to get permission to release the works on an Open Commons license — meaning that site visitors can reuse the materials. “One of our working principles is what we put on our site should be available for people to use downstream,” he said.
As for other works, “We identified what we call a core English canon,” he said. “If we can’t find a source for them, we will commission them.”
Among the fruits of that endeavor that are now being copyedited and soon will join the site: Avot d’Rav Natan, Sefer Hachinuch, and Tomer Devorah.
And it’s not just English translations he’s looking for. “We’re trying to serve some languages besides Hebrew and English,” he said. “I’m trying to track down a public domain translation of the Tanakh into Spanish. We have a Finnish translation. And our French translation just went up — it was done by a former chief rabbi of France.”
And as a computer scientist, Mr. Israel is working to automate some of the work that was the preserve of learned Torah scholars in earlier generations and more recently of manual indexers.
“When works on our site reference other works, we catch it and link it,” he said. We, in this case, being his algorithms. Now they’re updating the algorithms to capture less explicit linkages — the Hebrew equivalent of using “ibid.”
Because the site is open source, it’s easy for researchers to use the digitized texts for their own computer science research.
“I spoke with a Ph.D. student at Yale yesterday who is working on some automated methods to tag the argumentation in the Talmud,” he said. “As a first step, it’s intriguing. It gets to the question of whether a computer can understand some of the back-and-forth, and the playful idea of a computer being able to reason its way through the Talmud. We’re happy to support some of that.”
Sefaria has a partnership with Dicta, a group of computer science professors in Israel who work on problems in computer processing of natural language. One project they’re working on is automating the placement of nikudot, the Hebrew vowel markings, in texts that had been published without them.
“It’s possible to do a reasonable job with a computer,” he said. “Teachers using our texts in a classroom find it very helpful. It makes texts more accessible. It’s one of the applications of machine learning. If you feed it enough examples of texts with vowels it basically learns the patterns. It will not do as well as a human professional but certainly well enough. It gets in the high 90s percent correct — well enough to use in the classroom.” (You can try this program out for yourself at nakdan.dicta.org.il.)
Another Jewish digitization project, HebrewBooks.org, has PDFs of more than 50,000 Hebrew religious books online. Sefaria isn’t planning on matching that.
“The long tail is very long,” he said. “A lot of books are seldom referenced. We’ve been aiming for pretty much every book that someone would pick up in the beit midrash. The Tanakh with 50 commentaries, the Talmud with 20 commentaries, the Yerushalmi with a handful of commentaries, all of the midrashim.
“Once we get past that, rather than just going for breadth, we’re aiming more to have a really notable library of translations. Already we have Rashi on the Torah and much of Rashi on Tanakh in translation. We’re hoping to get Ibn Ezra and Ramban. Much of the core midrashim are already available in translation. That starts to open the core of rabbinic thought, the Jewish library, to people who couldn’t access it before.
“We’re going to spend a lot of time developing tools for teachers and learners. As technology keeps moving, there’s always more work to be done to keep things current. We’re starting to put thought in voice interfaces. What is Alexa’s place in this? How voice interfaces be helpful in a Torah conversation? It’s not an obvious question but it’s one it’s timely to ask.”
As for the site’s role in a Torah conversation, “we’re not pushing for a learner’s primary place of learning to be on Sefaria — though I wouldn’t be upset by it. There’s definitely an invaluable addition a computerized database can add to a printed page. The computer has the ability to hold and present a significant quantity of information. I would like to get to a place where someone learning a page of Talmud will feel they haven’t properly learned it until they look on our site and see parallel passages in the Yerushalmi and Tosefta.”
“There are people I know who learn primarily on Sefaria, but the more primary usage is as the reference shelf. It’s easier to look at a book for a long period of time. There’s something tactile about it that’s stable and reassuring. When there’s a question, a quote you want to track down, it’s much easier to have the phone or tablet next to you than to run to the shelf. And the phone gives you the ability to search, which you can’t do on the shelf. Using both the book and the device is the mode that’s most common in the beit midrash.
“It’s also helpful when you’re in transit. It gives you the ability to learn on the road.
“There’s another dynamic that is maybe less apparent, but also exciting: It gives access to people who don’t have a beit midrash. We got a letter from an aviator in the U.S. Air Force, thanking us for letting him learn while on duty in undisclosed locations.
“Also, non-Jews use us. We just spoke with a professor at Yale Divinity School who said the students are so happy that they can read these rabbinic commentaries on the Bible. It’s an area of fascination but it wasn’t accessible.
“It’s great that I can learn on my phone wherever I go, but it’s a trade-off. If I’m honest, I’ve got to admit I get distracted. A message pop ups and I’m chatting with someone. Or I closed the application and I’m on Facebook and I don’t remember how it happened. There’s something about the device that leads to a frenetic mindset. Personally, I’m comfortable with my primary focus being on a book. I find keeping my focus while looking at my phone for more than a minute is well nigh impossible. Technology trains us into a distractable mindset.
Mr. Israel admits that “there’s something bittersweet about putting it all online.” Nonetheless, he says, “It’s the call of the age. If people are going to be online, the Torah has to be online.
“The real Torah is learned in relationship, from a person’s mouth to a person’s ear. That’s why part of our mission is creating tools that can be used in the classroom, in conversations. We’re not trying to replace the magic or real transformative relationships and honest learning.”
This discussion of course isn’t unique to the Torah study.
“It’s really an example of the larger trends in human culture. We’re not the only ones going online. This isn’t the only place where things that were very physical are becoming more ephemeral, where barriers to communication and access are dropping, where things that used to be local and communal are becoming global. It’s upsetting to a lot of norms in a lot of ways.”
Sefaria is eager to help other cultures use its technology for organizing and displaying digitized texts.
“Our software is open. A couple of people tried to use it for Das Kapital by Karl Marx. We offered our support to a project coding the Buddhist scriptures. In the back of our head we would love for some group working with Islamic scripture to come along — we’ve done a lot of the graphic work of having the left to right text and right to left text be beautiful, and we’d love for more people to use it. We’re trying to get the word out.
“We have software that is free for the taking. We’re thinking of doing a couple projects ourselves so people can see what the potential is. Maybe Shakespeare or something from philosophy. It’s fun to think that the tools and norms of the beit midrash can echo into other spheres of learning.”
Who: Lev Israel, chief data officer for Sefaria.org
What: “Old Wine in New Vessels — Torah, Sefaria, and New Media”
When: Sunday, April 29, at 8 p.m.
Where: Congregation Rinat Yisrael, 389 W. Englewood Ave., Teaneck.
How much: Free
comments