BIBLE HISTORY DAILY

AI Unlocks Ancient Texts

How technology is helping fill in gaps in the Gilgamesh epic

AI and Ancient Texts: Plaque depicting Gilgamesh and Enkidu wrestling Humbaba. Courtesy Photo Companion to the Bible, Genesis.

While the Epic of Gilgamesh is one of the most famous ancient texts ever written and is available today in dozens of languages and editions, modern scholars have only succeeded in recovering around 70 percent of the original Akkadian epic. Since the first fragment of the text was identified in the late 1800s, Assyriologists (experts in cuneiform languages and cultures) have worked tirelessly to find new fragments to fill in the missing gaps in the text, a full copy of which has never been discovered. Launched in 2018, the AI project Fragmentarium is helping scholars identify not just fragments of Gilgamesh, but countless other ancient cuneiform texts.

FREE ebook: From Babylon to Baghdad. Ancient Iraq. Learn about Iraq and its cultural heritage. Download now.

First Name:*

Last Name:*

Email Address: * * Indicates a required field.

SUBMIT

If you don’t want to receive the Bible History Daily newsletter, uncheck this box.

Finding a Cuneiform Needle in a Fragmentary Haystack

Containing what is thought to be the earliest account of the great Flood, the Epic of Gilgamesh is a sequence of stories depicting the adventures of Gilgamesh, the semi-mythical king of Uruk. While scholars have been able to piece together most of the 12-tablet-long standard edition of the text, hundreds of lines are still unknown. Enter the Fragmentarium, a project of the Electronic Babylonian Library. By analyzing tens of thousands of known tablets and fragments, the Fragmentarium can quickly compare these to newly discovered fragments to find matches. To date, the project has transliterated nearly 30,000 cuneiform tablets, equaling approximately 400,000 lines of text.

Explaining to Bible History Daily how the project works, project researcher and University of Heidelberg Ph.D. candidate Luis Sáenz, said, “Researchers can input sequences of signs—such as those from the end of a broken line—to find tablets with matching sequences that might complete the text. Similarly, if only the beginning of lines are available, the system can search for matching signs across different lines of a text.” Since new fragments may contain parts of previously known text but also parts that had not yet been discovered, the program drastically increases the speed at which scholars can connect texts, empowering them to identify and add to the ever-growing database. The AI program has already added details to more than 100 lines of the Epic of Gilgamesh and thousands of lines from other texts, including Atra-Hasis, Enuma Elish, and the Great Shamash Hymn.

Replica of the Flood Tablet from the Epic of Gilgamesh. Courtesy Photo Companion to the Bible, Genesis.

Another major advantage of the Fragmentarium is its ability to utilize the work of past Assyriologists. Much of this work had remained locked away in the handwritten notes of scholars. By digitizing those notes and adding them to the Fragmentarium, the project allows users to access information that may have otherwise been lost. “A particularly interesting case was Carl Bezold’s legacy,” said Sáenz, “which had long been considered lost. Some scholars had mentioned seeing his transliterations, but their whereabouts were unknown. It was only after inquiring at the University of Heidelberg library and thoroughly searching that we rediscovered these unpublished notes, transliterations, and copies of tablets. This was a significant find for our project, underscoring the importance of building upon the foundational work of previous scholars rather than starting from scratch.”

There are at least half a million cuneiform tablets housed in museums and collections around the world, and more are discovered every year. With only a limited number of scholars capable of translating these millennia-old documents, the knowledge that they carry remains locked away. With projects such as the Fragmentarium, however, that may not be the case for long. “The primary impact of this project is acceleration. It’s crucial to emphasize that technology will never fully replace human scholarly work. Tools like AI and machine learning will, however, significantly speed up the processes of identifying and matching fragments, thereby aiding in the reconstruction of ancient literature.”

Related reading in Bible History Daily:

The Epic of Gilgamesh

Genesis and Gilgamesh

What Is Akkadian?

All-Access members, read more in the BAS Library:

Did Ecclesiastes Copy Gilgamesh?

Gilgamesh—Like You’ve Never Seen Him Before

 Gilgamesh: Hero, King, God and Striving Man