Ideas for Moving Voynich Research Forward

[pfeaster]

I see opportunities for progress exactly where there are odd statistics that lack an explanation.

I see definite opportunities for progress here too -- both in trying to come up with explanations that fit the odd statistics we have and in working out methods for enhancing or expanding upon those statistics.

On one hand, it seems important for us to keep trying to formulate scenarios or hypotheses that would be consistent with all the statistical oddities we already know about, even as their number continues to increase.  Of course there are plenty of hypotheses out there that might be able to explain one or two of them.  But I don't think I've yet read any prospective explanation, no matter how fanciful or far-fetched, that would simultaneously account for the low entropy, the rigid word morphology, the "language" differences (Currier A vs. B, etc.), the preferred and dispreferred word-break combinations discovered by Emma May Smith and Marco Ponzi, the networks of similar words (including co-occurrence patterns and frequencies based on edit distance from three base forms) found by Torsten Timm, the seven "eccentric habits" described by Tavi Stafford, the patterns of "rightwardness" and "downwardness" I've worked on myself, and the existence of so-called Jackson Sequences.  And I'm sure that's an incomplete list.  A productive question to pose here, I think, is: "What's the simplest writing system we could invent ourselves that would predictably display all of these weird characteristics?"  I know that's a daunting question, but approaching it from a creative design standpoint might make it easier to answer than thinking of it only hypothetically and abstractly.

On the other hand, there are further statistics I think might be helpful but don't know how we'd go about generating.  Thanks to Smith and Ponzi, we know that there's statistically significant patterning among word-break combinations.  But what about relationships among other parts of successive words that occupy the same "slots" as each other?  Consider the sequence lokar qokar okar okol ol.  Four words in a row contain ok, three words in a row contain ar, and two words in a row contain ol in, arguably, the same morphological "slot" as ar.  How often does each of those three things occur over a sequence of adjacent words, regardless of what the words are otherwise?  How common is the generic pattern qab ab ac c, where a, b, and c represent contrastive values, regardless of what those values are?  This kind of partial repetition is probably the most intuitively conspicuous pattern there is in Voynichese, but it's frustratingly hard to study on a large scale, apart from just calculating edit distances and so on.  If we could get some really robust and flexible tools for analyzing patterns of repetition, in a way that's sensitive to morphological "slots," I suspect they'd lead to some illuminating discoveries.