(28-09-2016, 01:43 PM)Diane Wrote: You are not allowed to view links. Register or Login to view....
If you have - for argument's sake - something like [glyph meaning 'city'] + [City name, abbreviated]+ [number]+['degrees' sign/letter] [N, S, W or E]..
would a linguist be able to determine which of the glyphs was being used as a number.. or as probably a number.. and which were abbreviations and so on?
...
This is both a good question and a good example. The reason I say it's a good example is because the structure of the VMS does, in fact, suggest that the document may be organized in a systematic way, with a possible emphasis on attributes.
By itself, it would be difficult for statistical analysis to decipher the individual parts of information coded in this way unless
- the ordering of the information were hierarchical or consistently ordered, and
- there were sufficient word-tokens of each kind for a pattern to be discernible.
In a simple-case scenario, if the text had a certain proportion of items like this
combined with narrative text, then it should be possible to pick out the two basic patterns and analyze how they differ. Then, if specific languages (or language groups with certain grammatical similarities) were compared against the groups of text, it may be possible to determine which are narrative and which are systematic, constructed, or heavily abbreviated. From there, it's a matter of unlocking one (usually the narrative form) and then using that information to try to decipher the more difficult parts (the systematic or abbreviated encodings).
A pattern like this would not be unusual. Hildegard von Bingen and others combined regular (unencoded) text with constructed text (made up words). In many diplomatic letters, only the sensitive parts were encoded. It's conceivable that the VMS contains a combination of narrative and constructed text in a self-devised alphabet.
In contrast, if the
entire manuscript were organized in the manner in which you describe, it would be a bigger statistical challenge, since there wouldn't be two kinds of patterns to compare against one another, but the system itself should be discernible if there is enough text to observe and "map" the patterns.
As for figuring out what the individual parts mean if the whole manuscript were systematically abbreviated or encoded, even if the building blocks can be broken apart, it's difficult to attach meaning unless you have references such as illustrations to affirm the process. We have to hope that the VMS text is related to the drawings and, if there's a system, that it's applied consistently.
Assuming the text and drawings are related, then statistics can look for self-similarity within individual sections in the hope that there would be more plant-related tokens in the plant sections and more cosmology tokens in the cosmological sections. Once this is determined, it becomes easier to assess the connections between various sections.
Patterns of frequency, self-similarity or hierarchy can be determined and charted reasonably well with statistical algorithms, but making sense of the data still requires a certain amount of "human computing" and insight.