Cipher or unknown language - historical perspective

[Jorge_Stolfi]

So say we have in front of us a medieval attempt to build an alphabet for a language. My suspicion is that a person well-acquainted with the relevant culture could still recognize the context in which this alphabet emerged.

Well, we know that the VMS alphabet was systematically constructed by combining a small set of simple strokes in almost every possible way. Thus the final letter shapes probably have no connection whatsoever to the cultural context of the Author, of the intended readers, of the language, or of the place where it was conceived. 

At the same time, [a person well-acquainted with the relevant culture would] probably already know the language or language family. [...] For any larger text written in and around Europe between let's say 1200-1500, I think this preliminary identification would take place in a matter of hours, not a century (and counting) like in the case of the VM.

I don't quite agree.  Starting with the "in and around Europe" bit.  However, for argument's sake, let's assume that the native range of Voynichese and the place where the script was created were indeed "in and around Europe between 1200-1500".

For the language to be identified, we first would need someone who can read that language as it was in that time frame.  There are plenty of such such people for languages that are still "alive", even peripheral ones like Armenian, Balkan Albanian, and Georgian.  But for languages that are extinct (like Caucasian Albanian, Bjarmian, Bulgar, Cuman, Crimean Gothic, Jassic, Khazar, Merya, Norn, Novgorodian, Pantesco, Polabian, Sabir, Sudovian, Ubikh, Yola, ..., and the Basque–Icelandic pidgin --- yes there was such a thing!) or nearly so (like Bats, Budukh, Cimbrian, Cypriot, Karaim, Kryz, Laz, Limousin, Livonian, Manx, Mocheno, Sami, Svan, Tsakonian, Udi, ...), the people with the necessary knowledge will be only a handful of linguists and language buffs, at best.

Then, that person would have to believe that Voynichese could be the language in question -- enough for him to be willing to make a serious effort at checking that possibility.  This is a big obstacle, because the "expert consensus", based on material and artistic evidence, is that Voynichese must be a West European language, Romance or South Germanic.  Knowing that, why would an expert on Udi or Khazar spend any time trying to see whether Voynichese is an exotic language well outside that range?

Even if the guess "maybe Sudovian" is correct, checking it will probably require a lot of work.  Voynichese is not just an original alphabet; it must be an original spelling system too.  The Alwan or Ubikh spelling that the expert knows may differ from the Voynichese one in many ways -- like the use of digraphs vs. single letters, the splitting and joining of affixes and auxiliary verbs, the handling of sandhi, allophones, tones, and epenthetic vowels, the use of abbreviations, etc..  

And it may be in a dialect with phonology and morphology different from those of the dialect he knows.  

How would he tackle the problem, considering that we still cannot guess the meaning of a single word in the whole book?

And then the Author may have had only an imperfect knowledge of the language, and made many errors of spelling and grammar.

And then Voynichese may be encrypted.  Even a simple trick, like writing the letters of each word backwards or inserting random nulls, would almost certainly prevent the identification of the language.

Just consider our difficulty in identifying the languages of the month names and of the "michton" marginalia.  Yet they are surely Western European, in the plain, written in the Latin alphabet, presumably with spelling typical for the language at the relevant time and place...

All the best, --jorge