A One-Page Ledger Method for Generating Voynich-Like Text

[Jorge_Stolfi]

No, there are far more words than just these four. They exist in a multidimensional network of dozens of related forms. The scribe doesn't choose between "otedy" and "oteedy" in isolation — he chooses among the entire visible pool of similar words:

Sorry, I don't see how that answers the objection below.  If anything, it makes it stronger:

If, after a suitable warm-up period, the words otedy and oteedy are equally frequent (as shown), and the mutation process can create ytedy from otedy, it should also create yteedy from oteedy. Then ytedy and yteedy should be equally frequent too.  But their ratio is only 1:6.

As I see it, the only ways your model would create the above counts are (1) the mutation of the prefix o->y is sensitive to whether the suffix is edy or eedy, or vice-versa, or (2) the seed text had those four words in those approximate skewed ratios (maybe no ytedy at all), and the mutation rules cannot create enough ytedy from otedy or from yteedy to raise the ytedy:yteedy ratio above 1:6.  Isn't that so?

Here is a larger sample for the ok-/k-/t-/ot- prefix group alone (not even including the y- or qo- variants):

Granted, compared to the patterns seen in typical texts in European languages, and even in Mandarin with tones, that table is more "Cartesian" ("closer to rank 1", with independent choices and/or mutations for prefix and suffix).  But it still has significant deviations from that model, which says that all the rows should be multiples of each other.  

For example,

  okain (144)  okar (129)
  otain (96)   otar (141)

There is no 1:1 relationship between two similar Voynich words. Each word exists in a network of variants.

Sorry again, I don't know what this means.

But within each prefix group, the ratios aren't uniform either — "okeey" (177) is more frequent than "okey" (63), "okaiin" (212) is more frequent than "okain" (144), reflecting which forms the scribe happened to use as sources more often.

Are you saying that the choice of the new starting point, whenever the "source pointer" for copying is reset, is also sensitive to the word at that point?  That is, the Scribe was more likely to restart the copying at an "okeey" than at an"okey"?  

Some cells are empty — "kaiir" (—), "tail" (—), "tam" (—). Not because a rule forbids them but because the scribe never happened to produce them.

That is the same as saying that the reason "tam" never occurred is that "tam" never occurred.

Take that "tam" as another example

  tam (0)   otam (47)
  tol (48)  otol (86)

Why didn't the mutation step ever produce a single "tam" from "otam" (or from any other nearby word, like tar (43) or kam (9))?

In fact, how come the frequency distribution of prefixes (or of suffixes) is so uneven, and so stable?  The mutation steps should gradually smooth out the distribution to some eigenvector of the transition matrix A (where A[i,j]) is the probability that prefix i mutates to prefix j)  That matrix must be highly skewed to produce such an uneven distribution...

That is what I mean by "frequent words are more likely to be selected as copying templates, generating more variants." Not that each word generates its variants at equal rates — but that the entire network of similar words feeds back on itself, with frequent forms generating more variants and rare forms generating fewer.

But that is the point. Once "otam" and "tol" and "tar" have become frequent, what is stopping them from mutating into "tam"?

An you still haven't answered the oldest objection.  The proposed algorithm for generating Voynichese-like gibberish text starts with
  Step 1. Generate a Voynichese-like gibberish seed text.
  Step 2. ... 
See the problem there?

All the best, --stolfi