Discovering Language: How AI Can Help Us Understand the Evolution of Old Languages

Throughout history, languages have developed and then died, evolved and then become extinct. This represents a vast chasm in the information that is available to us. While humans have been able to decode ancient languages that are unrelated to those available to us, we have not been able to track certain changes in the evolution of these languages. This is difficult given that languages change, symbols evolve to become unrecognizable, and of course, the history of human conquest means that entire languages have been wiped out by human actions. There is also the question of environmental changes, natural calamities, and the ways in which time tends to erode physically stored information.

Cracking ancient languages with AI reflects an unprecedented leap in historical and linguistic research. But this is not only a technological or scientific innovation. It also opens up many cultural avenues, the least interesting of which is the fact that never before in history has this kind of historical reach been available to us. It opens up not only technology, but also changes our capacity for studying humans themselves. Imagine, we can now access information about militaries, economies and ecologies which have left no physical traces behind.

In 2019, Jiaming Luo and their team of MIT researchers made headlines by using machine learning and AI technology to decipher ancient scripts. These had previously been decoded by Michael Ventris with some foundational work by Alice Kober, both in the late 20th century. This represents a huge milestone in the field.

Many indigenous cultures worldwide trace their origins to ancient languages like Sanskrit. By using Sanskrit as a reference, AI has been able to understand smaller, regional languages which have not yet been decoded. Through comparative analyses, AI can use what is called ‘structural sparsity’ to assist linguists and anthropologists to understand ancient cultures through what remains of these languages.

The greatest advantage that AI therefore offers is that it is able to work more effectively with lesser information to work off of than humans. Given its capacity to process large amounts of information, track patterns that are otherwise hard to detect and to have an almost unlimited capacity for the same, AI will serve as a tool to deal with the drudgery of data processing.

Sources:

1. https://toppandigital.com/translation-blog/using-ai-crack-ancient-languages/

2. https://restofworld.org/2022/indus-translation-ai-code-script/

3. https://www.linkedin.com/pulse/artificial-intelligence-key-decipher-ancient-languages-mona-fu/

Previous
Previous

From Morse Code to Machine Learning: Evolution of Language and Comms Technology

Next
Next

Certified Teacher: AI and the Future of Language Learning