Back

"Lost Languages: The Enigma of the World's Undeciphered Scripts" by Andrew Robinson

This was interesting enough. I wanted to know a bit about the languages from ancient civilizations and got what I came for. Apparently I am six years old because I thought the plethora of pictures were the most fun part, way more than the text.

The first part of the book is about finding and deciphering old texts in general. The second part describes the journey of some scripts humans can now interpret. The last part is about languages that are still a mystery (mostly because people can't find enough texts, because we destroyed them in the past and/or people can't find enough peoples who could interpret them or offer clues with related languages, because we murdered them all in the past).

Some parts bored me, some parts were very intriguing. There was more drama than I expected: quite some decipherers were extremely petty people who stole ideas, took credit, gatekept tablets, etc.

Somewhere there is a small discussion about why computers don't just crack these old languages for us. For most ancient languages, only a limited amount of texts are found, which is not enough material for a computer to see patterns in, so it doesn't have much helpful information to offer people working in the field. Humans are better at decoding with very few resources, with making connections with other languages, taking right guesses, having rare light bulb moments and such. This book is from 2002, so I don't know if that explanation is still relevant. But it did make me think of the shitty LLM/AI that is being pushed onto us in recent years. These systems have eaten almost all of the texts of the world by now, some are even starting to cannibalize their own nonsense, and still they can't compete with human language use and creativity.