I've always wondered.. how do they go about deciphering languages that are dead? Anyone know? :m:

Maybe you've heard of the famous "Rosetta Stone." The Egyptian hieroglyphic language remained untranslateable until a boy found this stone that contained just a small section, cracked in half, of Egyptian hieroglyphics with a Greek translation under it. From this small stone we were able to decipher the entire Egyptian written hieroglyphic languange. Amazing, isn't it? Until this time, you could say that Egyptian was a "dead" language. But this is not exactly true.

Sanskrit and ancient Greek and Latin are considered "dead" languages today because no one on earth uses them to communicate, so what you mean by "dead language" is different from this definition that we commonly use today.

Even without a "Rosetta Stone," I should think that computers in the near future - or perhaps even now! - could put together the relationships, connections, and uses of words or hieroglyphics (characters) between the uses (sentence patterns) of a socalled "dead language" to find matches that make sense and figure out their meaning.

The rosetta stone way is definetely the easiest way to understand a dead language... to find something that pretty much explains it all.

Valich, i like the way you thought about the computers and deciphering ways... a computer definetely could see how a sentence was structured, and figure out what it means by following the words exactly and seeing how they are placed... since all languages sort of blend together in certain places, they may even be able to look at other languages and see common relationships.

This is a little in-the-future... to program something that complex would take awhile, but it is definetely do-able. Wouldn't that be amazing?

I do not think a language could be deciphered without some contextual clues.

The order of the subject, main verb, & object of a sentence varies from one language to another. Noun/adjective & verb/adverb order varies from one language to another.

You need to find documents in contexts. From the format, you might be able to guess that a document is an invoice or a bill of sale. If it is found in a building used for smelting metal, you might be able to guess at words relating to tools or weapons. If it is found in some building used for storing grain or farm products, you might be able to guess at some words relating to farm products. Guessing or knowing that it is an invoice or bill of sale, you might be able to decide what words relate to numbers.

If the culture using the language is advanced enough to have illustrated books, you would have a fair chance of deciphering the language due to the pictures.

BTW: Does anybody know why some written languages are right-to-left, others vice versa, and others top to bottom? Somebody proposed the following, which makes sense to me.If a hammer and chisel were the first writing implements, the language would be written right to left. This is more convenient for a right handed person using a hammer & chisel.


If the first writing instrument was a pointed tool for inscribing a clay tablet or a quill pen using ink on a paper-like medium, the language would be left to right. This is best for a right handed person, who would smudge the words just written if he worked from right to left.


If the first writing implement was something like a small paint brush, top to bottom is most convenient.

Sanskrit and ancient Greek and Latin are considered "dead" languages today because no one on earth uses them to communicate, so what you mean by "dead language" is different from this definition that we commonly use today.


Sanskrit is taught to Indian children in schools from beginning to end. i don't have exact stats nor am i willing to look it up, but over 50% of indians can write in Sanskrit.

they look for the symbol that occurs most often. that is the E. if you find the next 5 most frequently occuring symbols you've got R L S T & N. everything else is done by using people who are really good at Wheel of Fortune. :)

Polywog: What you described is okay for solving simple substitution ciphers, assuming the language is English. I think the English letter frequency is ETOAIN SHRDLU or something similar, not ERSTLN. For Swahili or Russian, it is different.

If you have no knowledge of the language and no context, you are not going to decipher it. It should be possible to tell the difference between a language with an alphabet and a pictographic language like Chinese by the number of symbols used. More than that, forget it without some context or pictures associated with the symbols.

There is a Minoan Artefact from Crete called the Phaistos disk which is covered in printed symbols of some kind. This thing is about three and a half thousand years old, and is the only example of that particular kind of writing (if it is writing).
Despite decades of study, no conclusion can be reached about the meaning of the symbols. Without context and numerous examples to wiork with, it is literally impossible to decipher such writing (if that is what it is).
http://en.wikipedia.org/wiki/Phaistos_Disc

I do not think a language could be deciphered without some contextual clues.

Actually, I hadn't heard about the Rosetta stone before this.. It would be EXTREMELY tough to decipher a language without context (impossible for a single human), but could probably be done with supercomputers.

over 50% of indians can write in Sanskrit.

Do you mean Devenagarri script? Sanskrit is a spoken language, not really an alphabet.



eburacum45,
In the case of the Phaistos Disc, it would be pretty worthless, no? We don't even know which symbols are missing.

If somebody has a clue about how to do a job, it is possible that a computer mght be able to help due to superior number crunching capabilities. If nobody has a clue about how to start the job, a computer will not help.It would be EXTREMELY tough to decipher a language without context (impossible for a single human), but could probably be done with supercomputers.Is somebody posts a feasible method of attacking the problem, I might change my mind. Until then, forget translating a language without some context or external clues to help.

If nobody has a clue about how to start the job, a computer will not help.

Right, for now. When we make some strides in AI, I think the computer will be able to figure it out.


Is somebody posts a feasible method of attacking the problem, I might change my mind. Until then, forget translating a language without some context or external clues to help.

If you can crack the underlying patterns of human language and couple that with computer translation (which currently sucks) then it can be done. Till then..

Man, I lurve 'puters

The order of the subject, main verb, & object of a sentence varies from one language to another. Noun/adjective & verb/adverb order varies from one language to another.All languages do not have the same grammatical elements. Chinese does not have adjectives and adverbs except for, arguably, a very small set including words like "very." The words we translate as adjectives are "stative verbs," e.g., "to be large." "Large dog" is actually rendered as "being-large dog."

Xerxes: You are one of those who view computers as being magic rather than merely being very complex tools.If you can crack the underlying patterns of human language and couple that with computer translation (which currently sucks) then it can be done. Till then..Without some knowledge of the culture or other context, an unknown language just cannot be translated. To believe otherwise is the same as a belief in magic.

Contrary to the belief systems of many people, there are some problems which cannot be solved now and which will defy analysis forever more.

The Rosetta Stone is really quite famous as it is considered to be the origin for deciphering the Egyptian heiroglyphic language.

The inscriptions on the Rosetta Stone are repeated three times in three different languages, thus leading the way to the entire deciphering of Egyptian hierioglyphics:

1) Egyptiam hieroglyphics used by the ancient Egyptians are at the top of the stone.
2) Demotic translation, used by both Arabs and Egyptian was inscribed in the middle.
3) And then another translation in Greek was inscribed at the bottom.

It couldn't have been more perfect. For thousands of years, no one could decipher Egyptian heiroglyphics until the Rosetta Stone was found, but that was over a hundred years ago! By incorporating all the languages in the world in a computer today, I would find it hard to believe that a computer could not do the same thing today.