After a century failing to crack an ancient script, linguists turn to machines | Hacker News Reader

After a century failing to crack an ancient script, linguists turn to machines

theverge.com

141 pointsseycombi27 comments

The academic enmity in this field, between published authors, is huge and entertaining. These quotes from the article:

Wells’ beef with Witzel goes all the way back to his PhD dissertation on the Indus script, which Witzel tried to block, according to Wells. Later, while escorting Witzel through India, Wells would show him a PowerPoint presentation entitled “Ten reasons you don’t know what you’re talking about” while in the back of a cab.

and

“You would be better off getting medical advice from your garbage man than you would getting ideas about the Indus script from listening to Steve Farmer,”

I wonder if the entertainment value is an incentive for this sort of drama to continue.

Sayre's Law in action

Witzel comes across as a deeply racist person. I am told that he told an Indian person that "Hindus cant get a chair professorship to study Indology because they will be biased!". He has a long history of viciously attacking other scholars who try to criticize his work in academic fashion.

Having said that his book on world mythology is excellent.

>Witzel comes across as a deeply racist person. I am told that he told an Indian person that "Hindus cant get a chair professorship to study Indology because they will be biased!".

That's not racism.

I think you're on to something.

Indus Script is not currently in Unicode. There was a proposal in 1999 with about 250 code points. http://www.unicode.org/L2/L1999/n1959.pdf

I'd guess that factoid is less interesting than the story of why you know that. Care to explain?

I just looked it up on Google and Wikipedia. It's remarkable what you can learn with the right search terms. Mostly I just wanted to see all the glyphs and a Unicode code page is a good place to start.

til there is unicode for egyptian heiroglyph: https://en.wikipedia.org/wiki/Egyptian_Hieroglyphs_(Unicode_...

Also for cuneiform. And just about everything else.

Ideally all writing systems will eventually be in Unicode. It's a lot of work for historical systems though.

The problem, as TFA points out, is that 1) the inscriptions are short, on average 5 symbols, 2) the underlying language and even the language family is unknown.

Most (all?) decipherments have relied on having an idea of the language at least. What we have here are lots of seals with a bit of text, not paragraphs with real structure to analyze, maybe not even sentences.

Since they are seals, maybe they are mostly labels identifying the owner, or maybe some transactional label. In which case you will be hard pressed to extract a language from them. You probably will need longer examples to make progress.

If it does encode different languages in different places, like cuneiform, it might be worthwhile focusing on inscriptions found on the fringes, like in Iraq. Maybe some of them are actually Sumerian inscriptions or something like that.

The article suggests that the major impediment to solving the puzzle is funding, presumably for more archeological digs, and finding more clues.

I find it surprising that only in the past few years someone started to use statistical analysis (Markov, bigrams, trigrams) on this script. These are techniques that have been in use by cryptanalists for decades. Sounds like something very basic that you would do in the initial efforts to analyze any ancient script.

Not necessarily. IIRC, people trying to decipher Mayan glyphs with cryptanalytic methods were set on the wrong track and got nowhere, because a script is actually meant to communicate language, not to hide the message.

The headline reminds me of this

http://downlode.org/Etext/nine_billion_names_of_god.html

As a professional linguist, I think that Farmer et al. are probably right and it's almost surely not a "writing system" at all. (This is a term of art with a very specific sense.)

> (This is a term of art with a very specific sense.)

Would you care to enlighten us?

It's about productivity. If it's purely pictographic or idiographic, it can only encode precisely what the symbols mean. When cuneiform, for instance, went from its early pictographic/ideographic stage to a rhyme, rebus and initial stage, it went from being just a "tagging" system to an actual writing system, where anything that could be said could be encoded.

So a Google search tells me this is supposed to be the unicorn http://i.imgur.com/cdbIpEd.jpg

Could it simply be some sort of rope/wire wrapped around the animals head with some wooden/metal rod used for steering? There seems to be a saddle too?

Could it just be they only drew a single antler/horn?

Or it's a side image and one antler obscures the other

clearly a tachash

I was hoping this would be about the Voynich manuscript

https://en.wikipedia.org/wiki/Voynich_manuscript

I still like Xkcd's theory best:

https://xkcd.com/593/

Reply on news.ycombinator.com