Hi Anton

I am not sure if this is what you want, but given below are two links to a language-cum-mathematical cipher kaTapayadi system, usable only with Malayalam alphabet and originated in Kerala, India

http://www.dli.gov.in/rawdataupload/upload/insa/INSA_1/20005b66_273.pdf

http://en.wikipedia.org/wiki/Katapayadi_system

Narayanan Bhattathiri

Thank you. As I understand, this Katapayadi system maps arbitrary numerical sequences into meaningful text for easier memorizing. Definitively interesting, although not something that I'm looking for. Within the scope of my question, the output would be a meaningless text (perhaps using an artificial alphabet).

For example, a simple substitution cipher will produce a meaningless text, but it is not language-dependent, because using the same substitution rules you can well encrypt input texts in different languages (e.g. English, German, French etc.) that use the same Latin alphabet.

Yes, you are right; this is just a substitution of numbers with letters, and any language can use it; only thing is that if you don't know which language, you can't decipher it. But that is not a big problem now-a-days I imagine.

I mentioned it only because it is such an old one and well practiced by ancients. To read some poems, you will never think that the beautiful rhymes were codes; that is ingenious.

I have read some poems, may be 100yrs old or so, using what I call 'cross-language homonymy', ie. words with one meaning in one language and another in another language. In the poems I have read, Malayalam and Sanskrit were used. I don't know which is the oldest of these. May be they existed around 13th century when Malayalam poems started appearing in numbers and in new shape.

Another thing I have noticed are poems having Sanskrit words, but if the adjacent letters are rearranged, with out any change in their order, Malayalam words appear, and the whole meaning of the poem changes. This also has potential for cipher use. Our people at that time were really clever, especially with language.

Well, when you do a write up, you can mention these as potentials. May be you can even work-up some thing as the above two between, say German or English, or Russian-English.

Very long back, I read a story about the first English-Russian-English translation machine. It was given the sentence 'Spirit is willing; but the flesh is weak'. The machine translated it to Russian and then back to English, and what came out was 'The Vodka is good, but the meat is bad'! This is a very 'ciphersome' machine.

Narayanan

Антон, надо понимать, что найти достоверный документ времён 1330-1500 гг. н.э. почти невозможно. Написанные латинским алфавитом могут сохраниться только в Папской библиотеке в Ватикане. Но это практически недоступно.

Про документы (примеры шифрования) на кириллице полезно почитать МН Сперанского. Имейте в виду, что "кириллицу" изобрели после 1341г. н.э.

Попробуйте найти АНГЛИЙСКУЮ "Азбуку" хотя бы написанную латинским алфавитом на АНГЛИЙСКОМ же языке до 1550г. -- и поймёте о чём это пишу.

Dear Sergei,

Those examples are only for the sake of clarifying my question. I am not looking for Cyrillic, Latin (or any other) particular applications, but for the ciphering technique in itself, if any.

The question is not limited to the interval between 1330 - 1500 CE, I would be glad with any example from e.g. 12th century.

I am much surprised by your opinion that an authentic 1330-1500 CE document is hard to find. I believe there any many of those digitized and readily available on the Internet. Or do you mean that it is an encrypted document that is hard to find?

Israel:

Interesting! I never even thought of it in that way.

Although definitively not what I'm looking for. If I tried to answer my own question, then of course the first thing that would occur to me is stenography. Stenography is input language dependent. But stenography hardly can be deemed a cipher, it is rather a writing system. Other possible candidates would be steganographic approaches like those introduced by Trithemius, but that's later than 15th c.

This question of having characters in two different languages will not arise now because in Unicode representation you can represent all characters. Your problem is right when other encoding like ASCII are used.