+ Follow
image by pexels.com

Thousands of palm-leaf manuscripts and early paper manuscripts have been preserved in a precarious treasure all over the temples, village libraries and the old academic archives of South India, containing centuries of language, philosophy, science, religion, medicine, and memory. These are manuscripts in ancient versions of Tamil, Sanskrit, Grantha, and Vattezhuthu, which keep the intellectual pulse of the area alive. But day by day, they are softened away by the humidity and insects, and even the mere passage of time. The search for salvaging these doomed scripts has thus turned out to be one of the most pressing cultural endeavors in India, and the epicentre on this front is the emergence of digital preservation that is being driven by the latest technologies, such as AI and machine learning.

It is not just the physical deterioration of the manuscripts. The more serious problem is in the scripts. Most of these writings were composed in languages that are no longer spoken and systems of writing that are not well known to modern readers. Even a modern Tamil speaker, such as one, would have a hard time deciphering Vattezhuthu, the curved and old kind of Tamil that was written in palm-leaf manuscripts. Writings in old Sanskrit in the Grantha script even present an even more challenging problem since each character differs with the scribe or the area, or with the state of the leaf. By so doing, linguistic development as well as material decay are in line with preservation.

The issue of digital preservation in South India is, consequently, far more complicated than plain scanning. Every manuscript is a riddle that needs technological accuracy and linguistic skills. A lot of palm leaves are broken, and therefore, there are no lines of text; ink has been washed out of the fibers of the leaves, and words cannot be seen with the naked eye; fungus, smoke, and water leave stains which cause a distortion of characters. They can be ruined even by turning the pages. One digital archivist once wrote about working with such leaves that it feels like you are touching history and watching it disintegrate in your hands.

In order to deal with these issues, South Indian institutions have started to set up ambitious digital preservation initiatives. One of the most significant attempts is in Chennai and Thiruvananthapuram, where university research centers and cultural foundations have established special labs to save endangered manuscripts. These laboratories are a blend of the old and the new-age-breaking of the ancient books-revealing the reading of books that have not been read in hundreds of years.

One of the most interesting case studies is the activity of the Oriental Research Institute and Manuscripts Library in Thiruvananthapuram, which contains one of the largest collections of palm-leaf manuscripts in India. Through funding by the tech researchers and private foundations, the institute introduced a digital project that scans, ameliorates, and deciphers weak manuscripts with the help of AI and machine learning (ML). They are imprinted starting with multi-spectral imaging, which involves cameras recording varying wavelengths of light, such as ultraviolet and infrared, to display lines and characters under stains and age spots. Letters that are pale and blank to the human eye will be displayed on the computer screen in clear letter format.

After taking the pictures, the second difficulty arises: interpretation of the scripts. That is where machine learning comes in, transformative. Many sample images are fed to the system to train the AI models to identify ancient characters. Gradually, the model acquires the ability to identify minor differences between similar characters, to make predictions on missing letter parts, and suggest corrections. This is a breakthrough in the case of Grantha and old Tamil manuscripts. Machines now work in seconds to help in conducting a task that would have taken decades of scholarly training.

But AI does not just decipher. It also organizes. In the case of institutions that contain thousands of unclassified manuscripts, one of the most difficult aspects is cataloguing. Unless they are organized, the scholars cannot find texts about medicine, philosophy, astrology, literature, or even mathematics. AI-based cataloguing systems can scan text and find keywords based on metadata and categorize manuscripts by theme, approximate date, and subject. This produces digital libraries which are easily searchable by students, historians, linguists and researchers anywhere in the world.

An example of the methods followed in projects in Chennai, e.g. students at the University of Madras with assistance of independent manuscript foundations, are similar. Deep-learning models specifically trained on palm-leaf manuscripts of the temples of Tamil Nadu have been used by some laboratories, where many collections held by families have been found. Some of them collaborate with other international organizations to establish open-access archives, so that no manuscript is locked away or unknown to a limited number of managers.

The effect of such online activities is tremendous. First, they maintain knowledge which would be lost over time. Early versions of Ayurvedic medicine, astronomical arithmetic, moral teachings, and cultural histories exist in many manuscripts that allow historians to trace the development of the South Indian societies. Second, they rekindle dying scripts. Different scripts such as Vattezhuthu, Grantha and Modi are also receiving a resurgence of academic interest due to their increased accessibility by artificial intelligence. Third, there is the democratization of access in digital preservation. A student will now be able to access them anywhere in the world using the computer or a smartphone as opposed to traveling to a temple library and soliciting the delicate manuscripts, which cannot be exposed to light or human touch.

Nevertheless, the process is not over. Simple estimates suggest that in South India alone, there are millions of uncatalogued manuscripts, which are either wrapped in old cloth, or in metal trunks in the family residence. Several custodians are afraid to give them up because they have sentimental attachment and others lack historical significance of the documents in their possession. These manuscripts will keep on falling apart without the quick process of digitization.

In addition, AI technology is yet to overcome severely torn leaves, infrequent scripts, and untidy texts. Human scholars will still be needed to perform corrections, interpretation and contextualization of the things identified by the machines. Digital preservation does not mean the loss of professionalism; it is a means of multiplying it.

Ultimately, the fight against the loss of threatened South Indian manuscripts is a tussle between time, deterioration and oblivion. Yet it is also a story of hope. Due to the digital innovation, ancient wisdom is reborn. The scripts which were about to die are reborn on the burning screens. And in the silent rooms of Chennai and Thiruvananthapuram where the muffled wall of scanners mingles with the odiferousness of ancient palm leaves India is being slowly ferried into the future, one manuscript at a time.

References:

Discus