Digital cuneiforms: Updated tool expands access to ancient Hittite texts | Hacker News Reader

Digital cuneiforms: Updated tool expands access to ancient Hittite texts

phys.org

36 pointsardanarac18 comments

Better access for Hittie texts? I'm already deluged with nešili texts of the "Triple your barley yields with this one weird trick" type or, "Phrygian raiders don't want you to know this secret". Lately they've been trying to get me into some kind of currency scheme. I'll stick with bartering, thank you very much.

I'm thinking of just blocking all cuneiform texts.

> Lately they've been trying to get me into some kind of currency scheme.

If you mean coins, those are thought to be significantly more recent than the Hittites.

If you mean trade that is, for convenience, denominated in quantities of a reference commodity, that's much, much older. Everyone was doing that before anyone had noticed the Hittites.

They may be talking of this new, decentralized, open-sea, counterfeit-proof currency I've heard fishermen raving about. It's called cowrie. But some friends working in the mines think it's a joke and it will soon crash.

Silver in standardized amounts as a form of payment is attested before the Hittites are. (The same goes for grain, obviously.) Cowries as payment are from an entirely separate region, but presumably were no more difficult to produce than smelted metals were.

They even started to make some meme cowrie shells in China.

The tool url is not listed in TFA, but a simple google search results in the following:

https://www.hethport.uni-wuerzburg.de/TLHdig/

Why do journalists not reference the things they are reporting on these days?

The cynic in me says that it may be to avoid pushing the original source to the top of results, ensuring their secondary reporting gets more clicks. But the realist says that it's sheer laziness.

If we wanted to digitize all ancient texts, what would it take? How many are there? Starting at the beginning of written history ~5,200 years ago to ~2,000 years ago, how many texts do we have from each millenia or century? Are they accessible to be digitized?

What about earlier symbols and similar potential communication? For example, caves sometimes have marks that seems to have some symbolic meaning, such as slashes that seem like counts. How many are there to digitize?

I'm curious about the scale of it.

More than $5 million, probably less than $100 million, probably - would need a team of qualified experts to know what to photograph and record, a team trained to handle delicate and valuable artifacts, IT people, machine learning experts, gophers, grunts, and finally a team of people to work with institutions. It's not a super complex endeavor, logistically speaking.

I don't think it'll ever happen. Some institutions would resist being part of any third party scaled up attempt to record things, others would demand editorial control, and the most asinine petty politics would gum things up.

Specifically for cuneiform, there are multiple efforts, covering roughly the time period you mentioned. There is this post, [1], and [2].

There are probably around 1-2 million cuneiform tablets that have been found so far. Many of them complete, but even more of them as fragments. Those fragments mostly just sit around in store rooms. It's a giant puzzle few people even have enough knowledge to attempt.

1: https://cdli.mpiwg-berlin.mpg.de/

2: https://oracc.museum.upenn.edu/

Great, thanks. I've run across ORACC before, but I didn't know widespread digitization was part of their aim (i.e., I had the impression it was a database of artifacts digitized elsewhere).

> There are probably around 1-2 million cuneiform tablets that have been found so far. Many of them complete, but even more of them as fragments.

That's a pretty amazing number - that so many have survived. It seems like it would take too much luck, unless some library has endured this long.

Those tablets are indestructible once you expose them to fire. The records come from sacked cities. There's no shortage.

If you wanted to design a writing medium that enables archeology, it would be difficult to do better than clay tablets. Fired clay is great for records meant to last, being very fire and water resistant, while unfired clay was reusable if you just made it a bit wet again, great as reusable writing surfaces.

So many paper, parchment and papyrus records that were meant to be kept have been lost to fire and water over the millennia. Yet with clay tablets a fire not only struggles to harm the records that were meant to last, it turns all the temporary records, scratch pads and notes left around into long-lasting fired clay. Burn a Sumerian city to the ground and you create a snapshot of all their writing for archeologists to find millennia later

It’s not easy to access them for digitalisation. The museum which holds them, have to get them out of storage for you which involves a curator and his time. Also afaik many aren’t labelled extensive making it difficult to choose the ones you are looking for.

When your estimation is accurate, then there are likely hundred thousand fragments in simple labelled boxes, laying around in storages. To digitalise them would require a lot of money, the agreement of the museums and personnel.

...and more importantly, could we then leverage LLMs to fill in the blanks of ancient history?

What blanks, and how could LLMs accurately fill them?

You could use LLMs to give you rough passes at a translation of the texts probably, but they're pretty bad at catching the interesting/surprising details.

Reply on news.ycombinator.com