AI reveals multiple scribes behind Dead Sea Scrolls, challenging decades of scholarship

We can finally greet them through their handwriting
A researcher reflects on what it means to identify individual scribes in ancient texts through computational analysis.

Seventy years after their discovery, the Dead Sea Scrolls have surrendered a secret that human perception alone could never unlock. Researchers at the University of Groningen, guided by artificial intelligence, have found that the Great Isaiah Scroll was written not by one hand but by many — each scribe leaving a biomechanical signature invisible to the naked eye but legible to algorithms trained to measure the geometry of ancient letters. The finding reframes these sacred manuscripts not merely as texts, but as traces of a living community of scribes who shared knowledge, tradition, and craft across generations.

  • For decades, the apparent uniformity of the Dead Sea Scrolls' handwriting led scholars to a confident but incomplete conclusion: a single scribe had produced them.
  • No human eye could reasonably compare all five thousand instances of the letter 'aleph' in the Isaiah Scroll — and that limitation had quietly held scholarship in place for seventy years.
  • A team led by a professor of ancient Judaism and an AI specialist developed algorithms to measure the curvature, pressure, and spacing of individual characters at a microscopic level, finding patterns calligraphy had concealed.
  • The analysis revealed at least two scribes on the Isaiah Scroll whose styles were so similar they likely trained together, suggesting a scribal workshop rather than a solitary copyist.
  • The method can now be extended across the entire Qumran collection, promising to map networks of influence and connection woven into the production of humanity's oldest biblical texts.

Seventy years after their discovery in the caves of Qumran, the Dead Sea Scrolls have yielded a secret that human eyes alone could never quite see. Researchers at the University of Groningen, working with artificial intelligence, have determined that these ancient biblical texts were not the work of a single scribe — as scholars long believed based on their uniform handwriting — but of multiple hands whose individual signatures were hidden in the microscopic geometry of each letter.

The Great Isaiah Scroll contains the Hebrew letter aleph roughly five thousand times. No human could reasonably compare all five thousand instances by sight. Researchers led by Mladen Popović and computer scientist Lambert Schomaker developed algorithms to measure the biomechanical characteristics of each letter — curvature, pressure, spacing — at a microscopic level. The patterns that emerged told a different story than calligraphy had suggested.

Published in PLOS ONE under the project title "The Hands that Wrote the Bible," the study revealed at least two scribes at work on the Isaiah Scroll. More striking still, their styles were so similar that they likely trained together or shared a common scribal tradition — pointing not to isolated individuals but to a community, perhaps a workshop, where technique passed from one generation to the next.

What makes the work significant is not just the finding but the method. The computer does not see better than a human — it measures differently, processing data at a scale human perception cannot match. Schomaker described how digital imaging allows calculations at the micro level of individual characters, operating below the threshold of conscious handwriting variation, in the realm of biomechanical habit.

Popović offered a thought that captures the emotional weight of the discovery. "We will never know their names," he said of the ancient scribes. "But after seventy years of study, it seems we can finally greet them through their handwriting." The Dead Sea Scrolls, it turns out, are not just texts. They are traces of human hands — and now, for the first time, we can begin to tell one hand from another.

Seventy years after their discovery in the caves of Qumran, the Dead Sea Scrolls have yielded a secret that human eyes alone could never quite see. Researchers at the University of Groningen, working with artificial intelligence, have determined that these ancient biblical texts were not the work of a single scribe, as scholars had long believed based on their uniform handwriting. Instead, multiple hands wrote them—hands whose individual signatures were hidden in the microscopic geometry of each letter.

The Dead Sea Scrolls remain among the most important archaeological finds of the twentieth century. They contain the oldest known versions of the Hebrew Bible, along with previously unknown Jewish documents that illuminate life in the ancient world. For decades, paleographers examined them the traditional way: by eye, comparing letterforms and trying to spot variations in style. The uniformity they observed led to a straightforward conclusion—one scribe, one hand, one person working through these texts with meticulous care.

But the Great Isaiah Scroll, one of the most famous of the collection, contains the letter aleph—the first letter of the Hebrew alphabet—roughly five thousand times. No human could reasonably compare all five thousand instances by sight alone. This is where the computer enters. Researchers led by Mladen Popović, a professor of ancient Judaism at Groningen, and Lambert Schomaker, a computer scientist specializing in artificial intelligence, developed a method to measure the biomechanical characteristics of each letter at a microscopic level. They scanned the scroll digitally and programmed algorithms to analyze the curvature, pressure, and spacing of thousands of individual characters. The patterns that emerged told a different story than the one calligraphy had suggested.

The study, published in PLOS ONE and funded by the European Research Council under the project title "The Hands that Wrote the Bible," revealed that at least two scribes worked on the Isaiah Scroll. More striking still, these two scribes shared a remarkably similar writing style—so similar that it suggested they had trained together or come from the same scribal tradition. This finding opens a window onto the social world of ancient manuscript production. These were not isolated individuals copying texts in isolation. They were part of a community, a workshop perhaps, where knowledge and technique were transmitted from one generation to the next.

Popović described the discovery as emotionally resonant. "It opens a new window to the ancient world, which can reveal much more intricate connections between the scribes who produced the manuscripts," he said. The implications extend beyond the Isaiah Scroll. This computational approach can now be applied to other texts in the Qumran collection, potentially identifying individual scribes across multiple documents and revealing networks of connection and influence that shaped the production of these foundational religious texts.

What makes this work significant is not just the finding itself but the method. For seventy years, scholars had reached the limits of what the human eye could discern. The computer does not see better—it measures differently. It can process data at a scale that human perception cannot match, finding patterns in five thousand repetitions of a single letter. Schomaker explained the technical advantage plainly: digital imaging allows for calculations at the micro level of individual characters, measuring the precise curvature of each one. The result is a form of paleography that operates below the threshold of conscious handwriting variation, in the realm of biomechanical habit.

Popović offered a closing thought that captures something of what this research means. "We will never know their names," he said of the ancient scribes. "But after seventy years of study, it seems we can finally greet them through their handwriting." The Dead Sea Scrolls, in other words, are not just texts. They are traces of human hands, and now, for the first time, we can begin to distinguish one hand from another.

It opens a new window to the ancient world, which can reveal much more intricate connections between the scribes who produced the manuscripts
— Mladen Popović, professor of ancient Judaism at University of Groningen
After seventy years of study, it seems we can finally greet them through their handwriting
— Mladen Popović
A Conversa do Hearth Outra perspectiva sobre a história
Inventor

Why does it matter that multiple scribes wrote these scrolls rather than one? Doesn't the text stay the same either way?

Model

The text itself does stay the same, but the story behind it changes completely. If one scribe wrote it, you're looking at one person's labor, one person's choices. If multiple scribes wrote it, you're looking at a community—people who trained together, who shared techniques, who worked in the same tradition. That tells you something about how knowledge was preserved and transmitted in the ancient world.

Inventor

And the computer found something the human eye couldn't see for seventy years. How is that possible if the handwriting looks uniform to experts?

Model

Because the computer isn't looking at handwriting the way we do. It's measuring biomechanical patterns—the actual curve and pressure of five thousand individual letters. A human eye sees "this looks like one person's hand." A computer sees "the angle of this aleph differs by two degrees from that one, consistently, across hundreds of instances." Those tiny variations are invisible to us but meaningful to the algorithm.

Inventor

So the two scribes who wrote Isaiah had similar styles. Does that mean they were literally trained together?

Model

That's what the evidence suggests, but we can't be certain. They could have trained together, or learned from the same teacher, or come from the same scribal school. The point is that their writing styles are too similar to be coincidence. It points to a connection, a shared origin of some kind.

Inventor

What happens next? Can this method be used on other scrolls?

Model

Yes. This is really just the beginning. Now that they've proven the method works on Isaiah, they can apply it to the entire Qumran collection. You could potentially map out which scribes worked on which texts, identify individuals across multiple documents, and start to see networks of connection in ancient manuscript production. It's like being able to fingerprint the ancient world.

Inventor

But we'll never know who these people actually were.

Model

No. Their names are lost. But that's almost beside the point now. We can recognize them. We can say, "This is the hand of one person, and this is the hand of another." After two thousand years, that's a kind of greeting.

Quer a matéria completa? Leia o original em SAPO 24 ↗
Fale Conosco FAQ